search

SoftwareGift / FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019

Code for 3rd Place Solution in Face Anti-spoofing Attack Detection Challenge @ CVPR2019,model only 0.35M!!! 1.88ms(CPU)
Other
928 stars 282 forks source link

如何获取实际的预测结果啊？ （how to get predict result for one image?） #36

Open ApexPredator1 opened 5 years ago

ApexPredator1 commented 5 years ago

如何根据下面的代码的最后部分获取预测结果啊？ 比如最终的预测分数


# 加载模型
    model = models.fishnet150()

    # 加载模型文件的权重数据到model模型
    checkpoint = torch.load('checkpoints/pre-trainedModels/fishnet150_ckpt.tar')  # 加载指定的模型文件
    model.load_state_dict(checkpoint['state_dict'])  # 导入模型中的权重数据

    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")  # 设置计算硬件，CPU 或 GPU
    torch.cuda.set_device(device)
    # torch.backends.cudnn.benchmark = True
    model = model.module.to(device)
    model.eval()

    # 准备transform
    img_size = 224
    ratio = 224.0 / float(img_size)
    normalize = transforms.Normalize(mean=[0.14300402, 0.1434545, 0.14277956],
                                     std=[0.10050353, 0.100842826, 0.10034215])
    transform = transforms.Compose([transforms.Resize(int(224 * ratio)), transforms.CenterCrop(img_size),
                                    transforms.ToTensor(), normalize])

    # 读取单张人脸RGB图像进行预测
    img = Image.open("face1.jpg")
    img = img.convert('RGB')
    img = img.resize((224, 224))
    img = transform(img)

    img = np.array(img)  # 形状为(3, 224, 224)
    img = np.expand_dims(img, 0)  # 此时形状为(1, 3, 224, 224)

    input = torch.tensor(img, dtype=torch.float32, device=device)  
    output = model(input)  # output形状为[1, 1000]

    soft_output = torch.softmax(output, dim=-1)  # 对output中最后一个维度上的元素进行softmax处理，得到的形状依然是[1, 1000]
    preds = soft_output.to('cpu').detach().numpy()  # 转换为numpy数组，形状为(1, 1000)
jiagh2010 commented 5 years ago

@ApexPredator1 输入的数据是单个RGB人脸图，还是三张，RGB、深度图。。

ApexPredator1 commented 5 years ago

@ApexPredator1 输入的数据是单个RGB人脸图，还是三张，RGB、深度图。。 由上述代码可知，输入的数据是单张RGB人脸图像，它的numpy数组的形状是(1, 3, 224, 224)

jiagh2010 commented 5 years ago

对于RGB-D-IR数据，作者是分别对这三个数据进行训练的，只输入RGB应该是只是可见光下的结果预测吧，你试验成功了吗，如何验证的，效果如何？@ApexPredator1

ApexPredator1 commented 5 years ago

@jiagh2010 我上面的代码就是对单张RGB进行预测的结果，可是预测结果不是一个分数数值，而是一个元素数量为1000的一维向量，不知道怎么通过这个一维向量来判断当前图像是真是假？

例如：输出的一维向量 [3.15416983e-04, 1.60965152e-04, 2.36804801e-04, 9.14450866e-05, 8.80960761e-06, 5.39372741e-06, 1.01817541e-05, 3.35081932e-05, 5.67194438e-05, 1.89028124e-05, 1.98000795e-04, 4.65863850e-05, 6.02081709e-05, 6.33493182e-05, 4.29628817e-05, 4.67544451e-05, 5.97168946e-06, 9.08007041e-06, 4.14463684e-05, 1.96080910e-05, 5.75062313e-06, 2.42902403e-04, 1.56605020e-04, 2.71100667e-03, 1.32133160e-03, 9.32473849e-05, 2.05908189e-04, 1.37664319e-03, 7.19649717e-04, 2.61043697e-05, 3.18926723e-05, 2.42954065e-05, 4.50689869e-04, 1.08143744e-04, 1.54490517e-05, 1.02788836e-04, 1.37255760e-04, 6.26214605e-05, 4.14923197e-05, 3.32425188e-05, 5.09263737e-05, 4.34865651e-05, 4.58599970e-04, 2.63721868e-05, 1.64105531e-05, 1.55232483e-05, 7.84045478e-05, 3.18438033e-05, 1.75129055e-04, 2.53382663e-04, 1.73602210e-04, 3.41194915e-04, 2.63347902e-04, 5.49552096e-05, 3.36587691e-04, 1.55624861e-04, 6.19643019e-04, 2.00193055e-04, 5.38270979e-05, 4.23811543e-05, 2.02731055e-04, 6.05249152e-05, 6.46199987e-05, 1.40051474e-04, 7.08279767e-05, 1.97070185e-05, 8.09960766e-05, 3.02212546e-03, 6.66579072e-05, 2.50251236e-04, 8.81326792e-04, 4.97996400e-04, 1.83301314e-03, 1.18859136e-03, 2.92520766e-04, 5.63520752e-03, 7.11271559e-06, 5.90732088e-06, 2.52079681e-05, 1.75300670e-06, 6.98178519e-06, 6.19188177e-06, 4.15611294e-06, 7.12371038e-05, 4.43679513e-04, 1.72207128e-05, 1.61515491e-04, 1.04131732e-05, 2.15389100e-05, 9.99338517e-05, 9.82613565e-05, 1.97166610e-05, 4.59404037e-05, 6.51481878e-06, 3.52307507e-06, 1.75370333e-05, 1.68101687e-05, 2.71497647e-05, 4.90335078e-05, 1.33578375e-04, 2.99761959e-05, 1.62074903e-05, 5.90486779e-05, 3.68451816e-04, 1.56408598e-04, 1.86965874e-04, 1.68102968e-04, 6.69644156e-04, 2.05475750e-04, 4.17298288e-04, 1.56601891e-03, 1.25370425e-04, 5.05386670e-05, 4.29018764e-05, 7.42119038e-04, 1.00523474e-04, 4.27435676e-04, 1.98408336e-04, 5.93496254e-04, 7.18346913e-04, 1.35883344e-02, 7.09191518e-05, 1.64686484e-04, 4.45991327e-06, 3.74015372e-06, 4.02908790e-06, 3.04886125e-05, 9.25682798e-06, 3.12556258e-05, 7.65705954e-06, 4.59891671e-05, 7.32789995e-05, 1.63787536e-05, 4.17832052e-06, 4.51596361e-06, 7.19164609e-06, 5.29099225e-06, 9.06966852e-06, 3.18134312e-06, 4.48910987e-06, 1.88558552e-05, 1.09080411e-05, 1.08528752e-06, 2.66833085e-05, 2.38244011e-05, 2.12633640e-05, 5.00988390e-05, 9.89821856e-05, 4.98185000e-06, 7.56183945e-05, 1.01219899e-04, 2.80787499e-05, 1.29546324e-05, 2.42686365e-05, 9.77031777e-06, 2.72978032e-05, 4.28155145e-05, 4.57028182e-05, 1.82831485e-04, 4.84039811e-05, 8.37084281e-06, 6.41581664e-06, 7.62195714e-06, 1.00085599e-05, 8.70947260e-05, 4.43059434e-06, 2.14056922e-06, 2.46516574e-05, 4.62340950e-06, 6.11687301e-06, 1.38736077e-05, 4.90005050e-06, 1.87302999e-06, 1.06607649e-05, 2.25607691e-05, 6.23857704e-05, 9.00099349e-06, 3.19241826e-06, 7.57676025e-05, 1.82758777e-05, 8.32395017e-05, 1.34715418e-04, 6.35557590e-05, 1.85638870e-04, 1.52891716e-05, 3.91741487e-05, 2.58289379e-06, 6.03748849e-05, 8.27086260e-05, 1.55589310e-04, 4.09917311e-06, 1.15322306e-04, 1.08290496e-05, 1.97140489e-05, 3.08331346e-06, 1.40291435e-04, 8.23808659e-06, 4.68827893e-05, 1.45658405e-05, 5.67892450e-04, 5.32543854e-05, 1.26017421e-05, 4.38741199e-06, 1.18186363e-04, 8.34197417e-05, 4.98086192e-05, 3.20762229e-06, 1.52026914e-04, 1.16716792e-05, 9.66640437e-05, 1.67706476e-05, 1.25046881e-05, 4.97434894e-06, 2.29415000e-05, 8.39351378e-06, 1.02751495e-04, 2.80482654e-05, 7.29350222e-06, 1.86526431e-05, 3.78390803e-04, 3.71091155e-04, 4.92602376e-05, 5.31952765e-06, 3.06035945e-05, 4.52995209e-06, 1.80421448e-05, 4.89498525e-05, 1.73374465e-05, 1.03759694e-05, 1.12068970e-04, 1.18311204e-04, 1.00265017e-04, 2.66916832e-05, 8.02860595e-04, 2.45668853e-05, 2.33272131e-05, 1.15100411e-05, 8.27303575e-06, 4.57981951e-04, 9.48080415e-05, 1.30501130e-05, 2.00431663e-04, 4.39087762e-06, 2.98907853e-05, 2.90807584e-05, 3.33719545e-05, 8.12344588e-05, 7.16672148e-05, 1.70827225e-05, 1.32152752e-04, 3.73804447e-04, 1.75631430e-05, 6.79303994e-05, 7.84474287e-06, 1.79620969e-04, 3.16658407e-04, 4.50567837e-04, 3.92879228e-05, 5.45951370e-05, 1.59695002e-04, 1.74507113e-05, 1.60905911e-04, 4.10828034e-06, 1.01877959e-05, 8.20952300e-06, 6.42993518e-06, 3.76977405e-05, 4.28364656e-05, 5.59513137e-05, 3.77876917e-04, 4.45095829e-05, 2.58001473e-05, 1.42087549e-04, 2.96535960e-04, 1.09460361e-05, 4.07642838e-05, 7.51364350e-05, 8.38453765e-04, 1.27528782e-03, 5.82356181e-04, 3.65349610e-04, 2.79966247e-04, 3.17092345e-04, 1.35440874e-04, 8.72290169e-04, 1.39647047e-04, 3.60734848e-04, 7.92061328e-05, 2.87890347e-04, 1.29855573e-04, 1.45156300e-04, 6.66021442e-05, 1.36241593e-04, 1.51580434e-05, 7.74129148e-05, 1.08774599e-04, 4.16963267e-05, 3.77175049e-03, 9.70892652e-05, 2.65737821e-04, 2.57975160e-04, 1.72495013e-04, 1.68016355e-03, 4.63510623e-05, 1.58238778e-04, 3.54792457e-04, 1.90039619e-03, 2.69016833e-04, 1.65392164e-04, 3.31460824e-03, 1.00456248e-03, 4.90510487e-04, 8.20008718e-05, 4.97535693e-05, 9.59117606e-05, 5.12430677e-04, 3.40337283e-04, 1.79478331e-04, 6.03372719e-05, 3.66447406e-04, 1.02805061e-04, 2.13613603e-04, 3.16370279e-05, 3.16905393e-03, 1.40535145e-03, 1.60897529e-04, 5.75007798e-05, 3.18724815e-05, 3.38747996e-05, 2.88629020e-03, 2.93495978e-05, 1.60845539e-05, 5.17530680e-06, 5.92372344e-05, 8.85665140e-05, 1.92289099e-05, 3.28416645e-05, 3.82020662e-05, 4.79151913e-06, 5.61763500e-06, 6.90809611e-06, 4.08608639e-06, 6.35947390e-06, 2.97369661e-05, 8.88171257e-07, 8.69046005e-07, 1.53804206e-06, 1.71207660e-06, 8.75217756e-06, 7.52518872e-06, 6.18126114e-06, 3.55056654e-05, 3.91051144e-04, 3.80254169e-05, 9.42580664e-05, 2.71994450e-05, 2.18085843e-05, 1.21818630e-04, 4.54518668e-05, 1.62197583e-04, 2.03203366e-04, 1.43235375e-04, 3.54475429e-04, 6.57675337e-05, 1.62968819e-04, 1.50535279e-05, 2.28088229e-05, 2.50890807e-05, 5.10879945e-05, 1.48464867e-04, 2.45117280e-05, 7.38137023e-05, 9.98504984e-05, 2.87330535e-04, 6.18543068e-04, 1.22782029e-03, 2.33916522e-04, 3.89577501e-04, 2.54104030e-04, 6.13406883e-05, 1.92977008e-04, 3.34412471e-05, 2.47790322e-05, 2.70945049e-04, 1.00065030e-04, 9.20986786e-05, 1.96933703e-04, 1.62014636e-04, 6.02152286e-05, 7.42952907e-05, 2.80749482e-05, 1.27568503e-03, 5.80694650e-05, 2.46928452e-04, 1.03263056e-03, 1.73354973e-04, 3.04692861e-04, 1.65666838e-03, 2.24993724e-04, 5.24618827e-06, 4.24291538e-05, 9.65071813e-05, 1.01092759e-04, 2.19297362e-04, 1.09501079e-05, 8.64323636e-04, 9.63828279e-05, 2.31554295e-04, 1.72215747e-04, 1.78588030e-04, 8.44087481e-05, 8.56516053e-05, 4.29556276e-05, 8.93402670e-04, 1.64722081e-03, 9.79739204e-02, 1.43851867e-04, 2.75909406e-04, 1.36967101e-05, 5.63106369e-05, 5.97558355e-05, 2.69970857e-04, 6.29706847e-05, 1.08612679e-04, 6.53220704e-05, 2.75985309e-04, 1.92871121e-05, 3.04160967e-05, 1.76593443e-04, 1.94989471e-03, 1.79003473e-04, 2.38368101e-03, 2.94204419e-05, 4.02493861e-05, 8.45929244e-05, 2.04321099e-04, 3.93609842e-03, 6.32084114e-03, 8.50640790e-05, 2.38844031e-03, 9.97999450e-06, 9.41590173e-04, 1.07249230e-04, 4.87300749e-05, 1.72922024e-04, 1.13703054e-05, 1.17293475e-05, 1.94129290e-03, 5.68857009e-04, 1.01249585e-04, 1.03932216e-04, 6.82916632e-03, 2.23813375e-04, 7.80001506e-02, 1.21952311e-04, 3.73361981e-04, 1.34902157e-05, 5.81922650e-04, 9.38152720e-04, 3.28414002e-03, 1.64184754e-03, 4.80585935e-04, 4.02774567e-05, 5.95941783e-05, 7.67312362e-04, 1.44790116e-04, 2.41533667e-02, 2.47137305e-05, 9.21804894e-05, 1.11332658e-04, 3.82743630e-04, 1.00583700e-03, 3.02314671e-04, 1.69846408e-05, 4.19366639e-04, 1.19593074e-04, 1.39675755e-03, 2.15832421e-04, 1.92929601e-05, 1.20319564e-04, 7.09784899e-06, 8.96853089e-05, 3.61232123e-05, 2.30027782e-03, 6.82696409e-04, 4.60518495e-04, 6.24930486e-04, 7.28062878e-05, 5.04693562e-05, 1.17537802e-05, 6.68803841e-05, 2.96153248e-05, 5.10614249e-04, 9.01670719e-05, 1.30645838e-03, 1.50181039e-03, 8.48722357e-06, 2.35149389e-04, 5.78140935e-05, 7.63228978e-04, 1.34801073e-03, 1.09666253e-04, 2.02537893e-04, 6.45500331e-05, 1.65516825e-03, 2.67331809e-04, 3.47618261e-05, 4.41603879e-05, 2.52790254e-04, 7.48417224e-04, 9.50200920e-05, 4.75640828e-03, 1.88393024e-05, 4.99241833e-05, 4.29085456e-03, 9.14704215e-05, 4.15994946e-05, 3.32319942e-05, 3.40217550e-04, 1.18218502e-03, 2.90783471e-04, 5.46828123e-05, 3.32173877e-05, 5.55228849e-04, 1.88903330e-04, 1.71105974e-04, 1.93216873e-03, 1.76376756e-03, 1.11523941e-05, 1.17320764e-04, 5.32409795e-05, 3.87433538e-05, 1.60958771e-05, 1.07073436e-06, 2.81519635e-04, 1.99488670e-04, 5.47889322e-05, 2.98111205e-04, 1.21676987e-02, 1.84195873e-04, 4.06084146e-05, 3.18067934e-04, 3.72393319e-04, 1.47277387e-06, 1.61282369e-05, 3.03792744e-03, 1.63611421e-05, 1.78270668e-04, 9.18091740e-03, 2.19390859e-05, 4.02185724e-05, 8.79958330e-04, 5.81405067e-04, 1.52693698e-04, 7.69234553e-04, 1.73079912e-04, 1.10265973e-03, 3.35513287e-05, 5.80023276e-03, 6.50478061e-04, 2.56411886e-06, 7.78641243e-06, 2.55249324e-04, 9.06799105e-05, 3.11704644e-04, 9.06738060e-05, 9.42769460e-04, 2.73962032e-05, 3.31976009e-03, 6.38879865e-05, 6.31125120e-04, 6.57289784e-05, 2.69112406e-05, 1.02415637e-04, 1.99806236e-05, 5.30727302e-05, 6.64008898e-04, 3.56539298e-04, 1.24995800e-04, 4.72516476e-05, 1.69906765e-02, 4.52463311e-04, 1.09768425e-05, 2.81926536e-04, 1.64262048e-04, 1.81370205e-03, 4.58795927e-04, 1.68149790e-03, 1.01748719e-05, 1.04896713e-03, 1.12351065e-03, 3.66910899e-05, 2.37184504e-04, 2.82216002e-04, 3.18880484e-05, 4.48512752e-03, 1.29208260e-04, 3.20154504e-05, 1.94961212e-05, 2.22197355e-06, 5.32069011e-04, 1.13479572e-03, 6.19212107e-04, 6.00433443e-03, 1.20498426e-03, 2.15767195e-05, 9.23131593e-04, 4.17604024e-04, 1.88147224e-05, 5.66223229e-04, 7.71204592e-04, 1.99407543e-04, 8.58557760e-04, 6.78692432e-03, 1.63895811e-03, 2.19458691e-03, 1.53829972e-03, 6.27387053e-05, 3.11128632e-03, 2.30337988e-04, 1.40323546e-05, 7.13851159e-06, 1.35386646e-01, 5.24319010e-04, 4.93396983e-05, 1.21624791e-03, 1.16887502e-04, 2.21110633e-04, 1.13393727e-03, 6.76785348e-05, 1.04465134e-05, 2.20272457e-04, 5.97780636e-05, 1.81057199e-04, 5.47915231e-04, 4.08698470e-05, 7.49643295e-05, 5.83778042e-03, 8.73576282e-05, 6.40910026e-03, 2.14145575e-02, 1.63372388e-05, 4.38069110e-04, 8.59977372e-05, 1.56921524e-04, 1.75094334e-04, 5.20529691e-03, 5.42974849e-05, 6.94168871e-03, 8.91501713e-06, 6.17560218e-05, 4.48958046e-04, 6.25564207e-05, 7.02582984e-05, 7.38467264e-04, 1.61513570e-04, 2.42281913e-05, 2.01255079e-05, 1.73983008e-05, 1.16270019e-06, 2.40707421e-03, 4.20070101e-05, 1.55872898e-04, 7.93334853e-04, 3.22882806e-05, 4.59450996e-04, 1.08365763e-04, 6.63328756e-05, 5.90003547e-05, 5.43520146e-04, 1.88875681e-04, 2.37549084e-05, 4.44964244e-05, 1.22761261e-03, 1.10686873e-03, 1.58691104e-03, 2.21389477e-04, 1.28301152e-04, 2.29432437e-04, 3.41505976e-04, 4.52475064e-03, 4.43244091e-04, 8.87415736e-05, 7.92549472e-05, 1.94386506e-04, 1.17373029e-05, 3.11825784e-06, 1.69753039e-04, 2.27057864e-03, 1.62062774e-04, 2.07518795e-04, 2.26157776e-04, 1.91156333e-03, 3.81103437e-03, 2.23897041e-05, 1.27228315e-03, 2.11576140e-03, 1.98037436e-04, 7.85743396e-06, 1.64309749e-04, 1.95135959e-04, 9.01881413e-06, 6.33986610e-06, 3.39003309e-04, 9.79356537e-06, 8.12258222e-04, 1.02335634e-03, 2.25757295e-03, 3.94835079e-04, 1.87522564e-05, 5.17410319e-03, 1.51875999e-03, 4.30911267e-03, 2.18255409e-05, 5.09975303e-04, 2.53026234e-03, 2.04854435e-03, 2.38388209e-04, 4.16115625e-04, 7.17134681e-04, 3.06047914e-05, 1.40658102e-03, 2.12535888e-05, 3.23811313e-04, 4.08252090e-04, 1.32227706e-05, 7.28080568e-06, 1.33122772e-03, 1.44456681e-05, 1.15905852e-04, 2.32215189e-05, 1.30502821e-03, 4.65539139e-04, 1.39846490e-03, 1.48768871e-04, 8.68080122e-07, 2.41275906e-04, 8.86810740e-05, 5.23720119e-05, 2.78895877e-05, 4.93882108e-05, 6.09869712e-05, 1.12005242e-03, 1.23207981e-04, 2.00627874e-05, 1.56172042e-04, 8.42664958e-05, 7.11681132e-05, 1.24748272e-04, 9.79428441e-05, 2.77862346e-05, 9.56724107e-05, 3.85188068e-05, 7.73060128e-06, 2.08538520e-04, 1.08948123e-04, 1.70120946e-03, 2.14195717e-03, 1.23349193e-04, 6.72314723e-04, 1.51796703e-04, 1.18979391e-04, 6.90544257e-04, 1.45806372e-03, 8.38129927e-06, 1.08485040e-03, 2.28439458e-04, 7.81000290e-06, 1.05856033e-02, 4.38377727e-03, 1.24652244e-04, 5.44213399e-04, 3.14800331e-04, 4.30008367e-05, 7.16262221e-05, 1.88255362e-04, 4.20282886e-05, 2.97345570e-04, 4.34412016e-03, 6.51011127e-04, 1.71346276e-03, 3.93552566e-03, 9.02512547e-05, 5.89355659e-05, 1.48089632e-04, 8.23207884e-05, 3.85821513e-05, 1.21706992e-03, 3.55113239e-04, 2.25127256e-03, 7.72560132e-04, 3.44353961e-04, 1.59593986e-03, 2.60453875e-04, 2.91126162e-05, 4.57892820e-05, 2.80454638e-03, 3.53970187e-04, 2.84084344e-05, 2.99847470e-05, 2.47414049e-04, 1.62191736e-03, 1.65985097e-04, 3.05804570e-05, 9.68208013e-04, 5.26915828e-04, 3.33393837e-05, 7.24629208e-05, 3.96366202e-04, 2.51013320e-03, 2.43575596e-05, 1.28630537e-03, 5.08088815e-05, 1.42697891e-05, 8.21775291e-03, 2.62462976e-03, 6.73847580e-06, 4.14191018e-05, 8.78736028e-05, 4.85016359e-03, 3.68836249e-04, 9.87687963e-05, 5.50712168e-04, 8.87749629e-05, 3.11757857e-03, 8.65355469e-05, 1.40265211e-05, 1.43929174e-05, 3.77162251e-05, 1.27143330e-05, 4.11039218e-03, 2.88484825e-05, 1.58473628e-03, 3.18496302e-03, 1.36090186e-03, 4.47441067e-04, 4.56444715e-04, 3.68735287e-03, 5.11475722e-04, 1.59510746e-04, 5.47615091e-05, 1.01628043e-02, 3.20144929e-04, 1.64197427e-05, 1.44899304e-05, 2.60301924e-04, 1.56333146e-03, 1.68770237e-03, 3.52717645e-04, 2.21857626e-05, 1.85100303e-03, 1.03010377e-03, 3.14649037e-06, 6.46207700e-05, 9.37905279e-05, 1.10766594e-03, 1.47057683e-04, 3.28630296e-04, 2.16684733e-02, 5.75492522e-05, 7.65562872e-05, 3.79846286e-04, 2.18264700e-04, 1.85113255e-04, 3.74390278e-04, 3.41528619e-04, 3.51919189e-05, 8.61310764e-05, 4.12136615e-05, 9.85622100e-05, 5.91767275e-06, 1.46325887e-03, 1.15089887e-03, 1.14337650e-04, 3.38442187e-05, 3.96098522e-03, 8.46180119e-05, 2.43670365e-05, 5.14228304e-05, 1.73647166e-03, 8.00362977e-05, 2.08661775e-03, 2.40392430e-04, 1.72521177e-04, 1.38832314e-04, 1.65971101e-03, 8.62220259e-06, 5.00951319e-05, 2.98538507e-04, 7.35829235e-04, 8.83467583e-05, 1.43905991e-05, 3.48631671e-04, 3.86607862e-05, 6.28878421e-04, 2.04876313e-04, 3.04545265e-05, 8.40629873e-05, 5.47287380e-03, 1.45081766e-02, 4.04888717e-03, 6.53274357e-04, 5.63437259e-03, 2.74443789e-03, 4.48337873e-04, 2.50337715e-03, 7.43731973e-04, 1.16080791e-03, 4.02721162e-05, 8.18868139e-05, 1.66225036e-05, 5.30174075e-05, 2.31458363e-03, 1.38466514e-03, 2.91585049e-04, 1.99756803e-04, 1.54485941e-04, 1.32268702e-04, 8.21987633e-05, 3.18550708e-04, 4.43006604e-04, 2.75482802e-04, 1.03912440e-04, 1.06662817e-04, 3.86159658e-03, 5.31826913e-03, 4.53150569e-05, 2.61415524e-04, 4.33373405e-03, 4.22978745e-04, 2.38688939e-04, 1.56680908e-04, 3.60671373e-04, 1.28130522e-03, 1.12345594e-03, 2.59141583e-04, 3.13589589e-05, 1.83989585e-04, 4.87501820e-05, 9.47598717e-04, 1.05039013e-04, 1.09055883e-03, 3.94968738e-05, 1.07486371e-03, 4.87788534e-03, 4.21217224e-03, 4.01082169e-03, 1.10411714e-03, 1.04988692e-04, 7.82358169e-04, 3.79713974e-03, 8.26339237e-04, 2.15936510e-04, 4.59978776e-03, 1.30289540e-04, 4.40065560e-05, 1.92972031e-04, 5.96474019e-05, 2.02654424e-04, 1.45001322e-04, 1.99004222e-04, 1.02480903e-04, 4.47234232e-03, 9.89381573e-04, 2.03133607e-03, 1.43012541e-04, 4.82802898e-05, 4.63449443e-03, 3.66873792e-05, 3.28098744e-04, 6.63479659e-05, 4.32542838e-05, 8.00800080e-06, 1.70054918e-04, 1.36833347e-04, 5.26753538e-05, 6.92136178e-04, 1.81739684e-03, 4.29194537e-04, 1.48604944e-04, 3.46284301e-04, 4.06595122e-04, 9.76968477e-06, 7.77165813e-04, 1.71084423e-04, 2.45137839e-04, 4.96638066e-04, 7.92800856e-05, 1.89612561e-04, 3.52295647e-05, 8.73079553e-05, 9.26384091e-05, 6.26013643e-05, 1.22739584e-04, 1.02218101e-03, 6.59730518e-04]

jiagh2010 commented 5 years ago

@ApexPredator1 像我做人脸识别的时候提取的人脸特征

wmmbjut commented 5 years ago

您好，我想请问一下到底是怎样的训练顺序，里面有FishNet150, MobilenetV2, MoblieLitenet,FeatherNetA, FeatherNetB,这些很多的模型，作者说以fishnet和mobilenet作为预训练，是在这个基础上再继续训练吗，还有mobilelitenet，有点乱，可以回答一下吗？感激不尽！

HulkMaker commented 5 years ago

@jiagh2010 我上面的代码就是对单张RGB进行预测的结果，可是预测结果不是一个分数数值，而是一个元素数量为1000的一维向量，不知道怎么通过这个一维向量来判断当前图像是真是假？

例如：输出的一维向量 [3.15416983e-04, 1.60965152e-04, 2.36804801e-04, 9.14450866e-05, 8.80960761e-06, 5.39372741e-06, 1.01817541e-05, 3.35081932e-05, 5.67194438e-05, 1.89028124e-05, 1.98000795e-04, 4.65863850e-05, 6.02081709e-05, 6.33493182e-05, 4.29628817e-05, 4.67544451e-05, 5.97168946e-06, 9.08007041e-06, 4.14463684e-05, 1.96080910e-05, 5.75062313e-06, 2.42902403e-04, 1.56605020e-04, 2.71100667e-03, 1.32133160e-03, 9.32473849e-05, 2.05908189e-04, 1.37664319e-03, 7.19649717e-04, 2.61043697e-05, 3.18926723e-05, 2.42954065e-05, 4.50689869e-04, 1.08143744e-04, 1.54490517e-05, 1.02788836e-04, 1.37255760e-04, 6.26214605e-05, 4.14923197e-05, 3.32425188e-05, 5.09263737e-05, 4.34865651e-05, 4.58599970e-04, 2.63721868e-05, 1.64105531e-05, 1.55232483e-05, 7.84045478e-05, 3.18438033e-05, 1.75129055e-04, 2.53382663e-04, 1.73602210e-04, 3.41194915e-04, 2.63347902e-04, 5.49552096e-05, 3.36587691e-04, 1.55624861e-04, 6.19643019e-04, 2.00193055e-04, 5.38270979e-05, 4.23811543e-05, 2.02731055e-04, 6.05249152e-05, 6.46199987e-05, 1.40051474e-04, 7.08279767e-05, 1.97070185e-05, 8.09960766e-05, 3.02212546e-03, 6.66579072e-05, 2.50251236e-04, 8.81326792e-04, 4.97996400e-04, 1.83301314e-03, 1.18859136e-03, 2.92520766e-04, 5.63520752e-03, 7.11271559e-06, 5.90732088e-06, 2.52079681e-05, 1.75300670e-06, 6.98178519e-06, 6.19188177e-06, 4.15611294e-06, 7.12371038e-05, 4.43679513e-04, 1.72207128e-05, 1.61515491e-04, 1.04131732e-05, 2.15389100e-05, 9.99338517e-05, 9.82613565e-05, 1.97166610e-05, 4.59404037e-05, 6.51481878e-06, 3.52307507e-06, 1.75370333e-05, 1.68101687e-05, 2.71497647e-05, 4.90335078e-05, 1.33578375e-04, 2.99761959e-05, 1.62074903e-05, 5.90486779e-05, 3.68451816e-04, 1.56408598e-04, 1.86965874e-04, 1.68102968e-04, 6.69644156e-04, 2.05475750e-04, 4.17298288e-04, 1.56601891e-03, 1.25370425e-04, 5.05386670e-05, 4.29018764e-05, 7.42119038e-04, 1.00523474e-04, 4.27435676e-04, 1.98408336e-04, 5.93496254e-04, 7.18346913e-04, 1.35883344e-02, 7.09191518e-05, 1.64686484e-04, 4.45991327e-06, 3.74015372e-06, 4.02908790e-06, 3.04886125e-05, 9.25682798e-06, 3.12556258e-05, 7.65705954e-06, 4.59891671e-05, 7.32789995e-05, 1.63787536e-05, 4.17832052e-06, 4.51596361e-06, 7.19164609e-06, 5.29099225e-06, 9.06966852e-06, 3.18134312e-06, 4.48910987e-06, 1.88558552e-05, 1.09080411e-05, 1.08528752e-06, 2.66833085e-05, 2.38244011e-05, 2.12633640e-05, 5.00988390e-05, 9.89821856e-05, 4.98185000e-06, 7.56183945e-05, 1.01219899e-04, 2.80787499e-05, 1.29546324e-05, 2.42686365e-05, 9.77031777e-06, 2.72978032e-05, 4.28155145e-05, 4.57028182e-05, 1.82831485e-04, 4.84039811e-05, 8.37084281e-06, 6.41581664e-06, 7.62195714e-06, 1.00085599e-05, 8.70947260e-05, 4.43059434e-06, 2.14056922e-06, 2.46516574e-05, 4.62340950e-06, 6.11687301e-06, 1.38736077e-05, 4.90005050e-06, 1.87302999e-06, 1.06607649e-05, 2.25607691e-05, 6.23857704e-05, 9.00099349e-06, 3.19241826e-06, 7.57676025e-05, 1.82758777e-05, 8.32395017e-05, 1.34715418e-04, 6.35557590e-05, 1.85638870e-04, 1.52891716e-05, 3.91741487e-05, 2.58289379e-06, 6.03748849e-05, 8.27086260e-05, 1.55589310e-04, 4.09917311e-06, 1.15322306e-04, 1.08290496e-05, 1.97140489e-05, 3.08331346e-06, 1.40291435e-04, 8.23808659e-06, 4.68827893e-05, 1.45658405e-05, 5.67892450e-04, 5.32543854e-05, 1.26017421e-05, 4.38741199e-06, 1.18186363e-04, 8.34197417e-05, 4.98086192e-05, 3.20762229e-06, 1.52026914e-04, 1.16716792e-05, 9.66640437e-05, 1.67706476e-05, 1.25046881e-05, 4.97434894e-06, 2.29415000e-05, 8.39351378e-06, 1.02751495e-04, 2.80482654e-05, 7.29350222e-06, 1.86526431e-05, 3.78390803e-04, 3.71091155e-04, 4.92602376e-05, 5.31952765e-06, 3.06035945e-05, 4.52995209e-06, 1.80421448e-05, 4.89498525e-05, 1.73374465e-05, 1.03759694e-05, 1.12068970e-04, 1.18311204e-04, 1.00265017e-04, 2.66916832e-05, 8.02860595e-04, 2.45668853e-05, 2.33272131e-05, 1.15100411e-05, 8.27303575e-06, 4.57981951e-04, 9.48080415e-05, 1.30501130e-05, 2.00431663e-04, 4.39087762e-06, 2.98907853e-05, 2.90807584e-05, 3.33719545e-05, 8.12344588e-05, 7.16672148e-05, 1.70827225e-05, 1.32152752e-04, 3.73804447e-04, 1.75631430e-05, 6.79303994e-05, 7.84474287e-06, 1.79620969e-04, 3.16658407e-04, 4.50567837e-04, 3.92879228e-05, 5.45951370e-05, 1.59695002e-04, 1.74507113e-05, 1.60905911e-04, 4.10828034e-06, 1.01877959e-05, 8.20952300e-06, 6.42993518e-06, 3.76977405e-05, 4.28364656e-05, 5.59513137e-05, 3.77876917e-04, 4.45095829e-05, 2.58001473e-05, 1.42087549e-04, 2.96535960e-04, 1.09460361e-05, 4.07642838e-05, 7.51364350e-05, 8.38453765e-04, 1.27528782e-03, 5.82356181e-04, 3.65349610e-04, 2.79966247e-04, 3.17092345e-04, 1.35440874e-04, 8.72290169e-04, 1.39647047e-04, 3.60734848e-04, 7.92061328e-05, 2.87890347e-04, 1.29855573e-04, 1.45156300e-04, 6.66021442e-05, 1.36241593e-04, 1.51580434e-05, 7.74129148e-05, 1.08774599e-04, 4.16963267e-05, 3.77175049e-03, 9.70892652e-05, 2.65737821e-04, 2.57975160e-04, 1.72495013e-04, 1.68016355e-03, 4.63510623e-05, 1.58238778e-04, 3.54792457e-04, 1.90039619e-03, 2.69016833e-04, 1.65392164e-04, 3.31460824e-03, 1.00456248e-03, 4.90510487e-04, 8.20008718e-05, 4.97535693e-05, 9.59117606e-05, 5.12430677e-04, 3.40337283e-04, 1.79478331e-04, 6.03372719e-05, 3.66447406e-04, 1.02805061e-04, 2.13613603e-04, 3.16370279e-05, 3.16905393e-03, 1.40535145e-03, 1.60897529e-04, 5.75007798e-05, 3.18724815e-05, 3.38747996e-05, 2.88629020e-03, 2.93495978e-05, 1.60845539e-05, 5.17530680e-06, 5.92372344e-05, 8.85665140e-05, 1.92289099e-05, 3.28416645e-05, 3.82020662e-05, 4.79151913e-06, 5.61763500e-06, 6.90809611e-06, 4.08608639e-06, 6.35947390e-06, 2.97369661e-05, 8.88171257e-07, 8.69046005e-07, 1.53804206e-06, 1.71207660e-06, 8.75217756e-06, 7.52518872e-06, 6.18126114e-06, 3.55056654e-05, 3.91051144e-04, 3.80254169e-05, 9.42580664e-05, 2.71994450e-05, 2.18085843e-05, 1.21818630e-04, 4.54518668e-05, 1.62197583e-04, 2.03203366e-04, 1.43235375e-04, 3.54475429e-04, 6.57675337e-05, 1.62968819e-04, 1.50535279e-05, 2.28088229e-05, 2.50890807e-05, 5.10879945e-05, 1.48464867e-04, 2.45117280e-05, 7.38137023e-05, 9.98504984e-05, 2.87330535e-04, 6.18543068e-04, 1.22782029e-03, 2.33916522e-04, 3.89577501e-04, 2.54104030e-04, 6.13406883e-05, 1.92977008e-04, 3.34412471e-05, 2.47790322e-05, 2.70945049e-04, 1.00065030e-04, 9.20986786e-05, 1.96933703e-04, 1.62014636e-04, 6.02152286e-05, 7.42952907e-05, 2.80749482e-05, 1.27568503e-03, 5.80694650e-05, 2.46928452e-04, 1.03263056e-03, 1.73354973e-04, 3.04692861e-04, 1.65666838e-03, 2.24993724e-04, 5.24618827e-06, 4.24291538e-05, 9.65071813e-05, 1.01092759e-04, 2.19297362e-04, 1.09501079e-05, 8.64323636e-04, 9.63828279e-05, 2.31554295e-04, 1.72215747e-04, 1.78588030e-04, 8.44087481e-05, 8.56516053e-05, 4.29556276e-05, 8.93402670e-04, 1.64722081e-03, 9.79739204e-02, 1.43851867e-04, 2.75909406e-04, 1.36967101e-05, 5.63106369e-05, 5.97558355e-05, 2.69970857e-04, 6.29706847e-05, 1.08612679e-04, 6.53220704e-05, 2.75985309e-04, 1.92871121e-05, 3.04160967e-05, 1.76593443e-04, 1.94989471e-03, 1.79003473e-04, 2.38368101e-03, 2.94204419e-05, 4.02493861e-05, 8.45929244e-05, 2.04321099e-04, 3.93609842e-03, 6.32084114e-03, 8.50640790e-05, 2.38844031e-03, 9.97999450e-06, 9.41590173e-04, 1.07249230e-04, 4.87300749e-05, 1.72922024e-04, 1.13703054e-05, 1.17293475e-05, 1.94129290e-03, 5.68857009e-04, 1.01249585e-04, 1.03932216e-04, 6.82916632e-03, 2.23813375e-04, 7.80001506e-02, 1.21952311e-04, 3.73361981e-04, 1.34902157e-05, 5.81922650e-04, 9.38152720e-04, 3.28414002e-03, 1.64184754e-03, 4.80585935e-04, 4.02774567e-05, 5.95941783e-05, 7.67312362e-04, 1.44790116e-04, 2.41533667e-02, 2.47137305e-05, 9.21804894e-05, 1.11332658e-04, 3.82743630e-04, 1.00583700e-03, 3.02314671e-04, 1.69846408e-05, 4.19366639e-04, 1.19593074e-04, 1.39675755e-03, 2.15832421e-04, 1.92929601e-05, 1.20319564e-04, 7.09784899e-06, 8.96853089e-05, 3.61232123e-05, 2.30027782e-03, 6.82696409e-04, 4.60518495e-04, 6.24930486e-04, 7.28062878e-05, 5.04693562e-05, 1.17537802e-05, 6.68803841e-05, 2.96153248e-05, 5.10614249e-04, 9.01670719e-05, 1.30645838e-03, 1.50181039e-03, 8.48722357e-06, 2.35149389e-04, 5.78140935e-05, 7.63228978e-04, 1.34801073e-03, 1.09666253e-04, 2.02537893e-04, 6.45500331e-05, 1.65516825e-03, 2.67331809e-04, 3.47618261e-05, 4.41603879e-05, 2.52790254e-04, 7.48417224e-04, 9.50200920e-05, 4.75640828e-03, 1.88393024e-05, 4.99241833e-05, 4.29085456e-03, 9.14704215e-05, 4.15994946e-05, 3.32319942e-05, 3.40217550e-04, 1.18218502e-03, 2.90783471e-04, 5.46828123e-05, 3.32173877e-05, 5.55228849e-04, 1.88903330e-04, 1.71105974e-04, 1.93216873e-03, 1.76376756e-03, 1.11523941e-05, 1.17320764e-04, 5.32409795e-05, 3.87433538e-05, 1.60958771e-05, 1.07073436e-06, 2.81519635e-04, 1.99488670e-04, 5.47889322e-05, 2.98111205e-04, 1.21676987e-02, 1.84195873e-04, 4.06084146e-05, 3.18067934e-04, 3.72393319e-04, 1.47277387e-06, 1.61282369e-05, 3.03792744e-03, 1.63611421e-05, 1.78270668e-04, 9.18091740e-03, 2.19390859e-05, 4.02185724e-05, 8.79958330e-04, 5.81405067e-04, 1.52693698e-04, 7.69234553e-04, 1.73079912e-04, 1.10265973e-03, 3.35513287e-05, 5.80023276e-03, 6.50478061e-04, 2.56411886e-06, 7.78641243e-06, 2.55249324e-04, 9.06799105e-05, 3.11704644e-04, 9.06738060e-05, 9.42769460e-04, 2.73962032e-05, 3.31976009e-03, 6.38879865e-05, 6.31125120e-04, 6.57289784e-05, 2.69112406e-05, 1.02415637e-04, 1.99806236e-05, 5.30727302e-05, 6.64008898e-04, 3.56539298e-04, 1.24995800e-04, 4.72516476e-05, 1.69906765e-02, 4.52463311e-04, 1.09768425e-05, 2.81926536e-04, 1.64262048e-04, 1.81370205e-03, 4.58795927e-04, 1.68149790e-03, 1.01748719e-05, 1.04896713e-03, 1.12351065e-03, 3.66910899e-05, 2.37184504e-04, 2.82216002e-04, 3.18880484e-05, 4.48512752e-03, 1.29208260e-04, 3.20154504e-05, 1.94961212e-05, 2.22197355e-06, 5.32069011e-04, 1.13479572e-03, 6.19212107e-04, 6.00433443e-03, 1.20498426e-03, 2.15767195e-05, 9.23131593e-04, 4.17604024e-04, 1.88147224e-05, 5.66223229e-04, 7.71204592e-04, 1.99407543e-04, 8.58557760e-04, 6.78692432e-03, 1.63895811e-03, 2.19458691e-03, 1.53829972e-03, 6.27387053e-05, 3.11128632e-03, 2.30337988e-04, 1.40323546e-05, 7.13851159e-06, 1.35386646e-01, 5.24319010e-04, 4.93396983e-05, 1.21624791e-03, 1.16887502e-04, 2.21110633e-04, 1.13393727e-03, 6.76785348e-05, 1.04465134e-05, 2.20272457e-04, 5.97780636e-05, 1.81057199e-04, 5.47915231e-04, 4.08698470e-05, 7.49643295e-05, 5.83778042e-03, 8.73576282e-05, 6.40910026e-03, 2.14145575e-02, 1.63372388e-05, 4.38069110e-04, 8.59977372e-05, 1.56921524e-04, 1.75094334e-04, 5.20529691e-03, 5.42974849e-05, 6.94168871e-03, 8.91501713e-06, 6.17560218e-05, 4.48958046e-04, 6.25564207e-05, 7.02582984e-05, 7.38467264e-04, 1.61513570e-04, 2.42281913e-05, 2.01255079e-05, 1.73983008e-05, 1.16270019e-06, 2.40707421e-03, 4.20070101e-05, 1.55872898e-04, 7.93334853e-04, 3.22882806e-05, 4.59450996e-04, 1.08365763e-04, 6.63328756e-05, 5.90003547e-05, 5.43520146e-04, 1.88875681e-04, 2.37549084e-05, 4.44964244e-05, 1.22761261e-03, 1.10686873e-03, 1.58691104e-03, 2.21389477e-04, 1.28301152e-04, 2.29432437e-04, 3.41505976e-04, 4.52475064e-03, 4.43244091e-04, 8.87415736e-05, 7.92549472e-05, 1.94386506e-04, 1.17373029e-05, 3.11825784e-06, 1.69753039e-04, 2.27057864e-03, 1.62062774e-04, 2.07518795e-04, 2.26157776e-04, 1.91156333e-03, 3.81103437e-03, 2.23897041e-05, 1.27228315e-03, 2.11576140e-03, 1.98037436e-04, 7.85743396e-06, 1.64309749e-04, 1.95135959e-04, 9.01881413e-06, 6.33986610e-06, 3.39003309e-04, 9.79356537e-06, 8.12258222e-04, 1.02335634e-03, 2.25757295e-03, 3.94835079e-04, 1.87522564e-05, 5.17410319e-03, 1.51875999e-03, 4.30911267e-03, 2.18255409e-05, 5.09975303e-04, 2.53026234e-03, 2.04854435e-03, 2.38388209e-04, 4.16115625e-04, 7.17134681e-04, 3.06047914e-05, 1.40658102e-03, 2.12535888e-05, 3.23811313e-04, 4.08252090e-04, 1.32227706e-05, 7.28080568e-06, 1.33122772e-03, 1.44456681e-05, 1.15905852e-04, 2.32215189e-05, 1.30502821e-03, 4.65539139e-04, 1.39846490e-03, 1.48768871e-04, 8.68080122e-07, 2.41275906e-04, 8.86810740e-05, 5.23720119e-05, 2.78895877e-05, 4.93882108e-05, 6.09869712e-05, 1.12005242e-03, 1.23207981e-04, 2.00627874e-05, 1.56172042e-04, 8.42664958e-05, 7.11681132e-05, 1.24748272e-04, 9.79428441e-05, 2.77862346e-05, 9.56724107e-05, 3.85188068e-05, 7.73060128e-06, 2.08538520e-04, 1.08948123e-04, 1.70120946e-03, 2.14195717e-03, 1.23349193e-04, 6.72314723e-04, 1.51796703e-04, 1.18979391e-04, 6.90544257e-04, 1.45806372e-03, 8.38129927e-06, 1.08485040e-03, 2.28439458e-04, 7.81000290e-06, 1.05856033e-02, 4.38377727e-03, 1.24652244e-04, 5.44213399e-04, 3.14800331e-04, 4.30008367e-05, 7.16262221e-05, 1.88255362e-04, 4.20282886e-05, 2.97345570e-04, 4.34412016e-03, 6.51011127e-04, 1.71346276e-03, 3.93552566e-03, 9.02512547e-05, 5.89355659e-05, 1.48089632e-04, 8.23207884e-05, 3.85821513e-05, 1.21706992e-03, 3.55113239e-04, 2.25127256e-03, 7.72560132e-04, 3.44353961e-04, 1.59593986e-03, 2.60453875e-04, 2.91126162e-05, 4.57892820e-05, 2.80454638e-03, 3.53970187e-04, 2.84084344e-05, 2.99847470e-05, 2.47414049e-04, 1.62191736e-03, 1.65985097e-04, 3.05804570e-05, 9.68208013e-04, 5.26915828e-04, 3.33393837e-05, 7.24629208e-05, 3.96366202e-04, 2.51013320e-03, 2.43575596e-05, 1.28630537e-03, 5.08088815e-05, 1.42697891e-05, 8.21775291e-03, 2.62462976e-03, 6.73847580e-06, 4.14191018e-05, 8.78736028e-05, 4.85016359e-03, 3.68836249e-04, 9.87687963e-05, 5.50712168e-04, 8.87749629e-05, 3.11757857e-03, 8.65355469e-05, 1.40265211e-05, 1.43929174e-05, 3.77162251e-05, 1.27143330e-05, 4.11039218e-03, 2.88484825e-05, 1.58473628e-03, 3.18496302e-03, 1.36090186e-03, 4.47441067e-04, 4.56444715e-04, 3.68735287e-03, 5.11475722e-04, 1.59510746e-04, 5.47615091e-05, 1.01628043e-02, 3.20144929e-04, 1.64197427e-05, 1.44899304e-05, 2.60301924e-04, 1.56333146e-03, 1.68770237e-03, 3.52717645e-04, 2.21857626e-05, 1.85100303e-03, 1.03010377e-03, 3.14649037e-06, 6.46207700e-05, 9.37905279e-05, 1.10766594e-03, 1.47057683e-04, 3.28630296e-04, 2.16684733e-02, 5.75492522e-05, 7.65562872e-05, 3.79846286e-04, 2.18264700e-04, 1.85113255e-04, 3.74390278e-04, 3.41528619e-04, 3.51919189e-05, 8.61310764e-05, 4.12136615e-05, 9.85622100e-05, 5.91767275e-06, 1.46325887e-03, 1.15089887e-03, 1.14337650e-04, 3.38442187e-05, 3.96098522e-03, 8.46180119e-05, 2.43670365e-05, 5.14228304e-05, 1.73647166e-03, 8.00362977e-05, 2.08661775e-03, 2.40392430e-04, 1.72521177e-04, 1.38832314e-04, 1.65971101e-03, 8.62220259e-06, 5.00951319e-05, 2.98538507e-04, 7.35829235e-04, 8.83467583e-05, 1.43905991e-05, 3.48631671e-04, 3.86607862e-05, 6.28878421e-04, 2.04876313e-04, 3.04545265e-05, 8.40629873e-05, 5.47287380e-03, 1.45081766e-02, 4.04888717e-03, 6.53274357e-04, 5.63437259e-03, 2.74443789e-03, 4.48337873e-04, 2.50337715e-03, 7.43731973e-04, 1.16080791e-03, 4.02721162e-05, 8.18868139e-05, 1.66225036e-05, 5.30174075e-05, 2.31458363e-03, 1.38466514e-03, 2.91585049e-04, 1.99756803e-04, 1.54485941e-04, 1.32268702e-04, 8.21987633e-05, 3.18550708e-04, 4.43006604e-04, 2.75482802e-04, 1.03912440e-04, 1.06662817e-04, 3.86159658e-03, 5.31826913e-03, 4.53150569e-05, 2.61415524e-04, 4.33373405e-03, 4.22978745e-04, 2.38688939e-04, 1.56680908e-04, 3.60671373e-04, 1.28130522e-03, 1.12345594e-03, 2.59141583e-04, 3.13589589e-05, 1.83989585e-04, 4.87501820e-05, 9.47598717e-04, 1.05039013e-04, 1.09055883e-03, 3.94968738e-05, 1.07486371e-03, 4.87788534e-03, 4.21217224e-03, 4.01082169e-03, 1.10411714e-03, 1.04988692e-04, 7.82358169e-04, 3.79713974e-03, 8.26339237e-04, 2.15936510e-04, 4.59978776e-03, 1.30289540e-04, 4.40065560e-05, 1.92972031e-04, 5.96474019e-05, 2.02654424e-04, 1.45001322e-04, 1.99004222e-04, 1.02480903e-04, 4.47234232e-03, 9.89381573e-04, 2.03133607e-03, 1.43012541e-04, 4.82802898e-05, 4.63449443e-03, 3.66873792e-05, 3.28098744e-04, 6.63479659e-05, 4.32542838e-05, 8.00800080e-06, 1.70054918e-04, 1.36833347e-04, 5.26753538e-05, 6.92136178e-04, 1.81739684e-03, 4.29194537e-04, 1.48604944e-04, 3.46284301e-04, 4.06595122e-04, 9.76968477e-06, 7.77165813e-04, 1.71084423e-04, 2.45137839e-04, 4.96638066e-04, 7.92800856e-05, 1.89612561e-04, 3.52295647e-05, 8.73079553e-05, 9.26384091e-05, 6.26013643e-05, 1.22739584e-04, 1.02218101e-03, 6.59730518e-04]

你好，请问你得到这串结果的解读方式了吗？

ApexPredator1 commented 5 years ago

没有

发自我的iPhone

------------------ 原始邮件 ------------------ 发件人: HulKing notifications@github.com 发送时间: 2019年7月10日 22:01 收件人: SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019 FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019@noreply.github.com 抄送: ApexPredator1 576692510@qq.com, Mention mention@noreply.github.com 主题: 回复：[SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019] 如何获取实际的预测结果啊？ （how to get predict result for one image?） (#36)

@jiagh2010 我上面的代码就是对单张RGB进行预测的结果，可是预测结果不是一个分数数值，而是一个元素数量为1000的一维向量，不知道怎么通过这个一维向量来判断当前图像是真是假？

例如：输出的一维向量 [3.15416983e-04, 1.60965152e-04, 2.36804801e-04, 9.14450866e-05, 8.80960761e-06, 5.39372741e-06, 1.01817541e-05, 3.35081932e-05, 5.67194438e-05, 1.89028124e-05, 1.98000795e-04, 4.65863850e-05, 6.02081709e-05, 6.33493182e-05, 4.29628817e-05, 4.67544451e-05, 5.97168946e-06, 9.08007041e-06, 4.14463684e-05, 1.96080910e-05, 5.75062313e-06, 2.42902403e-04, 1.56605020e-04, 2.71100667e-03, 1.32133160e-03, 9.32473849e-05, 2.05908189e-04, 1.37664319e-03, 7.19649717e-04, 2.61043697e-05, 3.18926723e-05, 2.42954065e-05, 4.50689869e-04, 1.08143744e-04, 1.54490517e-05, 1.02788836e-04, 1.37255760e-04, 6.26214605e-05, 4.14923197e-05, 3.32425188e-05, 5.09263737e-05, 4.34865651e-05, 4.58599970e-04, 2.63721868e-05, 1.64105531e-05, 1.55232483e-05, 7.84045478e-05, 3.18438033e-05, 1.75129055e-04, 2.53382663e-04, 1.73602210e-04, 3.41194915e-04, 2.63347902e-04, 5.49552096e-05, 3.36587691e-04, 1.55624861e-04, 6.19643019e-04, 2.00193055e-04, 5.38270979e-05, 4.23811543e-05, 2.02731055e-04, 6.05249152e-05, 6.46199987e-05, 1.40051474e-04, 7.08279767e-05, 1.97070185e-05, 8.09960766e-05, 3.02212546e-03, 6.66579072e-05, 2.50251236e-04, 8.81326792e-04, 4.97996400e-04, 1.83301314e-03, 1.18859136e-03, 2.92520766e-04, 5.63520752e-03, 7.11271559e-06, 5.90732088e-06, 2.52079681e-05, 1.75300670e-06, 6.98178519e-06, 6.19188177e-06, 4.15611294e-06, 7.12371038e-05, 4.43679513e-04, 1.72207128e-05, 1.61515491e-04, 1.04131732e-05, 2.15389100e-05, 9.99338517e-05, 9.82613565e-05, 1.97166610e-05, 4.59404037e-05, 6.51481878e-06, 3.52307507e-06, 1.75370333e-05, 1.68101687e-05, 2.71497647e-05, 4.90335078e-05, 1.33578375e-04, 2.99761959e-05, 1.62074903e-05, 5.90486779e-05, 3.68451816e-04, 1.56408598e-04, 1.86965874e-04, 1.68102968e-04, 6.69644156e-04, 2.05475750e-04, 4.17298288e-04, 1.56601891e-03, 1.25370425e-04, 5.05386670e-05, 4.29018764e-05, 7.42119038e-04, 1.00523474e-04, 4.27435676e-04, 1.98408336e-04, 5.93496254e-04, 7.18346913e-04, 1.35883344e-02, 7.09191518e-05, 1.64686484e-04, 4.45991327e-06, 3.74015372e-06, 4.02908790e-06, 3.04886125e-05, 9.25682798e-06, 3.12556258e-05, 7.65705954e-06, 4.59891671e-05, 7.32789995e-05, 1.63787536e-05, 4.17832052e-06, 4.51596361e-06, 7.19164609e-06, 5.29099225e-06, 9.06966852e-06, 3.18134312e-06, 4.48910987e-06, 1.88558552e-05, 1.09080411e-05, 1.08528752e-06, 2.66833085e-05, 2.38244011e-05, 2.12633640e-05, 5.00988390e-05, 9.89821856e-05, 4.98185000e-06, 7.56183945e-05, 1.01219899e-04, 2.80787499e-05, 1.29546324e-05, 2.42686365e-05, 9.77031777e-06, 2.72978032e-05, 4.28155145e-05, 4.57028182e-05, 1.82831485e-04, 4.84039811e-05, 8.37084281e-06, 6.41581664e-06, 7.62195714e-06, 1.00085599e-05, 8.70947260e-05, 4.43059434e-06, 2.14056922e-06, 2.46516574e-05, 4.62340950e-06, 6.11687301e-06, 1.38736077e-05, 4.90005050e-06, 1.87302999e-06, 1.06607649e-05, 2.25607691e-05, 6.23857704e-05, 9.00099349e-06, 3.19241826e-06, 7.57676025e-05, 1.82758777e-05, 8.32395017e-05, 1.34715418e-04, 6.35557590e-05, 1.85638870e-04, 1.52891716e-05, 3.91741487e-05, 2.58289379e-06, 6.03748849e-05, 8.27086260e-05, 1.55589310e-04, 4.09917311e-06, 1.15322306e-04, 1.08290496e-05, 1.97140489e-05, 3.08331346e-06, 1.40291435e-04, 8.23808659e-06, 4.68827893e-05, 1.45658405e-05, 5.67892450e-04, 5.32543854e-05, 1.26017421e-05, 4.38741199e-06, 1.18186363e-04, 8.34197417e-05, 4.98086192e-05, 3.20762229e-06, 1.52026914e-04, 1.16716792e-05, 9.66640437e-05, 1.67706476e-05, 1.25046881e-05, 4.97434894e-06, 2.29415000e-05, 8.39351378e-06, 1.02751495e-04, 2.80482654e-05, 7.29350222e-06, 1.86526431e-05, 3.78390803e-04, 3.71091155e-04, 4.92602376e-05, 5.31952765e-06, 3.06035945e-05, 4.52995209e-06, 1.80421448e-05, 4.89498525e-05, 1.73374465e-05, 1.03759694e-05, 1.12068970e-04, 1.18311204e-04, 1.00265017e-04, 2.66916832e-05, 8.02860595e-04, 2.45668853e-05, 2.33272131e-05, 1.15100411e-05, 8.27303575e-06, 4.57981951e-04, 9.48080415e-05, 1.30501130e-05, 2.00431663e-04, 4.39087762e-06, 2.98907853e-05, 2.90807584e-05, 3.33719545e-05, 8.12344588e-05, 7.16672148e-05, 1.70827225e-05, 1.32152752e-04, 3.73804447e-04, 1.75631430e-05, 6.79303994e-05, 7.84474287e-06, 1.79620969e-04, 3.16658407e-04, 4.50567837e-04, 3.92879228e-05, 5.45951370e-05, 1.59695002e-04, 1.74507113e-05, 1.60905911e-04, 4.10828034e-06, 1.01877959e-05, 8.20952300e-06, 6.42993518e-06, 3.76977405e-05, 4.28364656e-05, 5.59513137e-05, 3.77876917e-04, 4.45095829e-05, 2.58001473e-05, 1.42087549e-04, 2.96535960e-04, 1.09460361e-05, 4.07642838e-05, 7.51364350e-05, 8.38453765e-04, 1.27528782e-03, 5.82356181e-04, 3.65349610e-04, 2.79966247e-04, 3.17092345e-04, 1.35440874e-04, 8.72290169e-04, 1.39647047e-04, 3.60734848e-04, 7.92061328e-05, 2.87890347e-04, 1.29855573e-04, 1.45156300e-04, 6.66021442e-05, 1.36241593e-04, 1.51580434e-05, 7.74129148e-05, 1.08774599e-04, 4.16963267e-05, 3.77175049e-03, 9.70892652e-05, 2.65737821e-04, 2.57975160e-04, 1.72495013e-04, 1.68016355e-03, 4.63510623e-05, 1.58238778e-04, 3.54792457e-04, 1.90039619e-03, 2.69016833e-04, 1.65392164e-04, 3.31460824e-03, 1.00456248e-03, 4.90510487e-04, 8.20008718e-05, 4.97535693e-05, 9.59117606e-05, 5.12430677e-04, 3.40337283e-04, 1.79478331e-04, 6.03372719e-05, 3.66447406e-04, 1.02805061e-04, 2.13613603e-04, 3.16370279e-05, 3.16905393e-03, 1.40535145e-03, 1.60897529e-04, 5.75007798e-05, 3.18724815e-05, 3.38747996e-05, 2.88629020e-03, 2.93495978e-05, 1.60845539e-05, 5.17530680e-06, 5.92372344e-05, 8.85665140e-05, 1.92289099e-05, 3.28416645e-05, 3.82020662e-05, 4.79151913e-06, 5.61763500e-06, 6.90809611e-06, 4.08608639e-06, 6.35947390e-06, 2.97369661e-05, 8.88171257e-07, 8.69046005e-07, 1.53804206e-06, 1.71207660e-06, 8.75217756e-06, 7.52518872e-06, 6.18126114e-06, 3.55056654e-05, 3.91051144e-04, 3.80254169e-05, 9.42580664e-05, 2.71994450e-05, 2.18085843e-05, 1.21818630e-04, 4.54518668e-05, 1.62197583e-04, 2.03203366e-04, 1.43235375e-04, 3.54475429e-04, 6.57675337e-05, 1.62968819e-04, 1.50535279e-05, 2.28088229e-05, 2.50890807e-05, 5.10879945e-05, 1.48464867e-04, 2.45117280e-05, 7.38137023e-05, 9.98504984e-05, 2.87330535e-04, 6.18543068e-04, 1.22782029e-03, 2.33916522e-04, 3.89577501e-04, 2.54104030e-04, 6.13406883e-05, 1.92977008e-04, 3.34412471e-05, 2.47790322e-05, 2.70945049e-04, 1.00065030e-04, 9.20986786e-05, 1.96933703e-04, 1.62014636e-04, 6.02152286e-05, 7.42952907e-05, 2.80749482e-05, 1.27568503e-03, 5.80694650e-05, 2.46928452e-04, 1.03263056e-03, 1.73354973e-04, 3.04692861e-04, 1.65666838e-03, 2.24993724e-04, 5.24618827e-06, 4.24291538e-05, 9.65071813e-05, 1.01092759e-04, 2.19297362e-04, 1.09501079e-05, 8.64323636e-04, 9.63828279e-05, 2.31554295e-04, 1.72215747e-04, 1.78588030e-04, 8.44087481e-05, 8.56516053e-05, 4.29556276e-05, 8.93402670e-04, 1.64722081e-03, 9.79739204e-02, 1.43851867e-04, 2.75909406e-04, 1.36967101e-05, 5.63106369e-05, 5.97558355e-05, 2.69970857e-04, 6.29706847e-05, 1.08612679e-04, 6.53220704e-05, 2.75985309e-04, 1.92871121e-05, 3.04160967e-05, 1.76593443e-04, 1.94989471e-03, 1.79003473e-04, 2.38368101e-03, 2.94204419e-05, 4.02493861e-05, 8.45929244e-05, 2.04321099e-04, 3.93609842e-03, 6.32084114e-03, 8.50640790e-05, 2.38844031e-03, 9.97999450e-06, 9.41590173e-04, 1.07249230e-04, 4.87300749e-05, 1.72922024e-04, 1.13703054e-05, 1.17293475e-05, 1.94129290e-03, 5.68857009e-04, 1.01249585e-04, 1.03932216e-04, 6.82916632e-03, 2.23813375e-04, 7.80001506e-02, 1.21952311e-04, 3.73361981e-04, 1.34902157e-05, 5.81922650e-04, 9.38152720e-04, 3.28414002e-03, 1.64184754e-03, 4.80585935e-04, 4.02774567e-05, 5.95941783e-05, 7.67312362e-04, 1.44790116e-04, 2.41533667e-02, 2.47137305e-05, 9.21804894e-05, 1.11332658e-04, 3.82743630e-04, 1.00583700e-03, 3.02314671e-04, 1.69846408e-05, 4.19366639e-04, 1.19593074e-04, 1.39675755e-03, 2.15832421e-04, 1.92929601e-05, 1.20319564e-04, 7.09784899e-06, 8.96853089e-05, 3.61232123e-05, 2.30027782e-03, 6.82696409e-04, 4.60518495e-04, 6.24930486e-04, 7.28062878e-05, 5.04693562e-05, 1.17537802e-05, 6.68803841e-05, 2.96153248e-05, 5.10614249e-04, 9.01670719e-05, 1.30645838e-03, 1.50181039e-03, 8.48722357e-06, 2.35149389e-04, 5.78140935e-05, 7.63228978e-04, 1.34801073e-03, 1.09666253e-04, 2.02537893e-04, 6.45500331e-05, 1.65516825e-03, 2.67331809e-04, 3.47618261e-05, 4.41603879e-05, 2.52790254e-04, 7.48417224e-04, 9.50200920e-05, 4.75640828e-03, 1.88393024e-05, 4.99241833e-05, 4.29085456e-03, 9.14704215e-05, 4.15994946e-05, 3.32319942e-05, 3.40217550e-04, 1.18218502e-03, 2.90783471e-04, 5.46828123e-05, 3.32173877e-05, 5.55228849e-04, 1.88903330e-04, 1.71105974e-04, 1.93216873e-03, 1.76376756e-03, 1.11523941e-05, 1.17320764e-04, 5.32409795e-05, 3.87433538e-05, 1.60958771e-05, 1.07073436e-06, 2.81519635e-04, 1.99488670e-04, 5.47889322e-05, 2.98111205e-04, 1.21676987e-02, 1.84195873e-04, 4.06084146e-05, 3.18067934e-04, 3.72393319e-04, 1.47277387e-06, 1.61282369e-05, 3.03792744e-03, 1.63611421e-05, 1.78270668e-04, 9.18091740e-03, 2.19390859e-05, 4.02185724e-05, 8.79958330e-04, 5.81405067e-04, 1.52693698e-04, 7.69234553e-04, 1.73079912e-04, 1.10265973e-03, 3.35513287e-05, 5.80023276e-03, 6.50478061e-04, 2.56411886e-06, 7.78641243e-06, 2.55249324e-04, 9.06799105e-05, 3.11704644e-04, 9.06738060e-05, 9.42769460e-04, 2.73962032e-05, 3.31976009e-03, 6.38879865e-05, 6.31125120e-04, 6.57289784e-05, 2.69112406e-05, 1.02415637e-04, 1.99806236e-05, 5.30727302e-05, 6.64008898e-04, 3.56539298e-04, 1.24995800e-04, 4.72516476e-05, 1.69906765e-02, 4.52463311e-04, 1.09768425e-05, 2.81926536e-04, 1.64262048e-04, 1.81370205e-03, 4.58795927e-04, 1.68149790e-03, 1.01748719e-05, 1.04896713e-03, 1.12351065e-03, 3.66910899e-05, 2.37184504e-04, 2.82216002e-04, 3.18880484e-05, 4.48512752e-03, 1.29208260e-04, 3.20154504e-05, 1.94961212e-05, 2.22197355e-06, 5.32069011e-04, 1.13479572e-03, 6.19212107e-04, 6.00433443e-03, 1.20498426e-03, 2.15767195e-05, 9.23131593e-04, 4.17604024e-04, 1.88147224e-05, 5.66223229e-04, 7.71204592e-04, 1.99407543e-04, 8.58557760e-04, 6.78692432e-03, 1.63895811e-03, 2.19458691e-03, 1.53829972e-03, 6.27387053e-05, 3.11128632e-03, 2.30337988e-04, 1.40323546e-05, 7.13851159e-06, 1.35386646e-01, 5.24319010e-04, 4.93396983e-05, 1.21624791e-03, 1.16887502e-04, 2.21110633e-04, 1.13393727e-03, 6.76785348e-05, 1.04465134e-05, 2.20272457e-04, 5.97780636e-05, 1.81057199e-04, 5.47915231e-04, 4.08698470e-05, 7.49643295e-05, 5.83778042e-03, 8.73576282e-05, 6.40910026e-03, 2.14145575e-02, 1.63372388e-05, 4.38069110e-04, 8.59977372e-05, 1.56921524e-04, 1.75094334e-04, 5.20529691e-03, 5.42974849e-05, 6.94168871e-03, 8.91501713e-06, 6.17560218e-05, 4.48958046e-04, 6.25564207e-05, 7.02582984e-05, 7.38467264e-04, 1.61513570e-04, 2.42281913e-05, 2.01255079e-05, 1.73983008e-05, 1.16270019e-06, 2.40707421e-03, 4.20070101e-05, 1.55872898e-04, 7.93334853e-04, 3.22882806e-05, 4.59450996e-04, 1.08365763e-04, 6.63328756e-05, 5.90003547e-05, 5.43520146e-04, 1.88875681e-04, 2.37549084e-05, 4.44964244e-05, 1.22761261e-03, 1.10686873e-03, 1.58691104e-03, 2.21389477e-04, 1.28301152e-04, 2.29432437e-04, 3.41505976e-04, 4.52475064e-03, 4.43244091e-04, 8.87415736e-05, 7.92549472e-05, 1.94386506e-04, 1.17373029e-05, 3.11825784e-06, 1.69753039e-04, 2.27057864e-03, 1.62062774e-04, 2.07518795e-04, 2.26157776e-04, 1.91156333e-03, 3.81103437e-03, 2.23897041e-05, 1.27228315e-03, 2.11576140e-03, 1.98037436e-04, 7.85743396e-06, 1.64309749e-04, 1.95135959e-04, 9.01881413e-06, 6.33986610e-06, 3.39003309e-04, 9.79356537e-06, 8.12258222e-04, 1.02335634e-03, 2.25757295e-03, 3.94835079e-04, 1.87522564e-05, 5.17410319e-03, 1.51875999e-03, 4.30911267e-03, 2.18255409e-05, 5.09975303e-04, 2.53026234e-03, 2.04854435e-03, 2.38388209e-04, 4.16115625e-04, 7.17134681e-04, 3.06047914e-05, 1.40658102e-03, 2.12535888e-05, 3.23811313e-04, 4.08252090e-04, 1.32227706e-05, 7.28080568e-06, 1.33122772e-03, 1.44456681e-05, 1.15905852e-04, 2.32215189e-05, 1.30502821e-03, 4.65539139e-04, 1.39846490e-03, 1.48768871e-04, 8.68080122e-07, 2.41275906e-04, 8.86810740e-05, 5.23720119e-05, 2.78895877e-05, 4.93882108e-05, 6.09869712e-05, 1.12005242e-03, 1.23207981e-04, 2.00627874e-05, 1.56172042e-04, 8.42664958e-05, 7.11681132e-05, 1.24748272e-04, 9.79428441e-05, 2.77862346e-05, 9.56724107e-05, 3.85188068e-05, 7.73060128e-06, 2.08538520e-04, 1.08948123e-04, 1.70120946e-03, 2.14195717e-03, 1.23349193e-04, 6.72314723e-04, 1.51796703e-04, 1.18979391e-04, 6.90544257e-04, 1.45806372e-03, 8.38129927e-06, 1.08485040e-03, 2.28439458e-04, 7.81000290e-06, 1.05856033e-02, 4.38377727e-03, 1.24652244e-04, 5.44213399e-04, 3.14800331e-04, 4.30008367e-05, 7.16262221e-05, 1.88255362e-04, 4.20282886e-05, 2.97345570e-04, 4.34412016e-03, 6.51011127e-04, 1.71346276e-03, 3.93552566e-03, 9.02512547e-05, 5.89355659e-05, 1.48089632e-04, 8.23207884e-05, 3.85821513e-05, 1.21706992e-03, 3.55113239e-04, 2.25127256e-03, 7.72560132e-04, 3.44353961e-04, 1.59593986e-03, 2.60453875e-04, 2.91126162e-05, 4.57892820e-05, 2.80454638e-03, 3.53970187e-04, 2.84084344e-05, 2.99847470e-05, 2.47414049e-04, 1.62191736e-03, 1.65985097e-04, 3.05804570e-05, 9.68208013e-04, 5.26915828e-04, 3.33393837e-05, 7.24629208e-05, 3.96366202e-04, 2.51013320e-03, 2.43575596e-05, 1.28630537e-03, 5.08088815e-05, 1.42697891e-05, 8.21775291e-03, 2.62462976e-03, 6.73847580e-06, 4.14191018e-05, 8.78736028e-05, 4.85016359e-03, 3.68836249e-04, 9.87687963e-05, 5.50712168e-04, 8.87749629e-05, 3.11757857e-03, 8.65355469e-05, 1.40265211e-05, 1.43929174e-05, 3.77162251e-05, 1.27143330e-05, 4.11039218e-03, 2.88484825e-05, 1.58473628e-03, 3.18496302e-03, 1.36090186e-03, 4.47441067e-04, 4.56444715e-04, 3.68735287e-03, 5.11475722e-04, 1.59510746e-04, 5.47615091e-05, 1.01628043e-02, 3.20144929e-04, 1.64197427e-05, 1.44899304e-05, 2.60301924e-04, 1.56333146e-03, 1.68770237e-03, 3.52717645e-04, 2.21857626e-05, 1.85100303e-03, 1.03010377e-03, 3.14649037e-06, 6.46207700e-05, 9.37905279e-05, 1.10766594e-03, 1.47057683e-04, 3.28630296e-04, 2.16684733e-02, 5.75492522e-05, 7.65562872e-05, 3.79846286e-04, 2.18264700e-04, 1.85113255e-04, 3.74390278e-04, 3.41528619e-04, 3.51919189e-05, 8.61310764e-05, 4.12136615e-05, 9.85622100e-05, 5.91767275e-06, 1.46325887e-03, 1.15089887e-03, 1.14337650e-04, 3.38442187e-05, 3.96098522e-03, 8.46180119e-05, 2.43670365e-05, 5.14228304e-05, 1.73647166e-03, 8.00362977e-05, 2.08661775e-03, 2.40392430e-04, 1.72521177e-04, 1.38832314e-04, 1.65971101e-03, 8.62220259e-06, 5.00951319e-05, 2.98538507e-04, 7.35829235e-04, 8.83467583e-05, 1.43905991e-05, 3.48631671e-04, 3.86607862e-05, 6.28878421e-04, 2.04876313e-04, 3.04545265e-05, 8.40629873e-05, 5.47287380e-03, 1.45081766e-02, 4.04888717e-03, 6.53274357e-04, 5.63437259e-03, 2.74443789e-03, 4.48337873e-04, 2.50337715e-03, 7.43731973e-04, 1.16080791e-03, 4.02721162e-05, 8.18868139e-05, 1.66225036e-05, 5.30174075e-05, 2.31458363e-03, 1.38466514e-03, 2.91585049e-04, 1.99756803e-04, 1.54485941e-04, 1.32268702e-04, 8.21987633e-05, 3.18550708e-04, 4.43006604e-04, 2.75482802e-04, 1.03912440e-04, 1.06662817e-04, 3.86159658e-03, 5.31826913e-03, 4.53150569e-05, 2.61415524e-04, 4.33373405e-03, 4.22978745e-04, 2.38688939e-04, 1.56680908e-04, 3.60671373e-04, 1.28130522e-03, 1.12345594e-03, 2.59141583e-04, 3.13589589e-05, 1.83989585e-04, 4.87501820e-05, 9.47598717e-04, 1.05039013e-04, 1.09055883e-03, 3.94968738e-05, 1.07486371e-03, 4.87788534e-03, 4.21217224e-03, 4.01082169e-03, 1.10411714e-03, 1.04988692e-04, 7.82358169e-04, 3.79713974e-03, 8.26339237e-04, 2.15936510e-04, 4.59978776e-03, 1.30289540e-04, 4.40065560e-05, 1.92972031e-04, 5.96474019e-05, 2.02654424e-04, 1.45001322e-04, 1.99004222e-04, 1.02480903e-04, 4.47234232e-03, 9.89381573e-04, 2.03133607e-03, 1.43012541e-04, 4.82802898e-05, 4.63449443e-03, 3.66873792e-05, 3.28098744e-04, 6.63479659e-05, 4.32542838e-05, 8.00800080e-06, 1.70054918e-04, 1.36833347e-04, 5.26753538e-05, 6.92136178e-04, 1.81739684e-03, 4.29194537e-04, 1.48604944e-04, 3.46284301e-04, 4.06595122e-04, 9.76968477e-06, 7.77165813e-04, 1.71084423e-04, 2.45137839e-04, 4.96638066e-04, 7.92800856e-05, 1.89612561e-04, 3.52295647e-05, 8.73079553e-05, 9.26384091e-05, 6.26013643e-05, 1.22739584e-04, 1.02218101e-03, 6.59730518e-04]

你好，请问你得到这串结果的解读方式了吗？

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

baoyu-yuan commented 5 years ago

@ApexPredator1 作者是采用了特征层1024维的前两个作为分类概率，所以同样你取前两个看哪个大呗

zhaotun commented 5 years ago

@baorave 能否结合本例详细说说

baoyu-yuan commented 5 years ago

@zhaotun 上面之所以1024维的值都差不多是因为测试数据没去掉人脸背景，如果和训练数据一样去掉了背景　得到的分数会集中在前两个　

mactiendinh commented 4 years ago

@baorave Do you verify accuracy when remove the background ?

baoyu-yuan commented 4 years ago

yes you can use penetration to remove background

---原始邮件--- 发件人: "mactiendinh"<notifications@github.com> 发送时间: 2019年12月25日(周三) 中午12:32 收件人: "SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019"<FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019@noreply.github.com>; 抄送: "Mention"<mention@noreply.github.com>;"baorave"<798976541@qq.com>; 主题: Re: [SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019] 如何获取实际的预测结果啊？ （how to get predict result for one image?） (#36)

@baorave Do you verify accuracy when remove the background ?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

mactiendinh commented 4 years ago

@baorave Can you share code or tool remove background from face image for me? Thank you so much

baoyu-yuan commented 4 years ago

you can find code of prnet in github

---原始邮件--- 发件人: "mactiendinh"<notifications@github.com> 发送时间: 2019年12月25日(周三) 下午3:07 收件人: "SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019"<FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019@noreply.github.com>; 抄送: "Mention"<mention@noreply.github.com>;"baorave"<798976541@qq.com>; 主题: Re: [SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019] 如何获取实际的预测结果啊？ （how to get predict result for one image?） (#36)

@baorave Can you share code or tool remove background from face image for me? Thank you so much

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

mactiendinh commented 4 years ago

prnet don't fit with embedded device, beacause prnet run only highly with GPU. I choose FeatherNet for run embedded device cause small size trained model. Can you suggest something else for remove background?

baoyu-yuan commented 4 years ago

If you utilize only the rgb image, maybe you can try with face landmarks.  For example, to fit an ellipse over the face area.

---原始邮件--- 发件人: "mactiendinh"<notifications@github.com> 发送时间: 2019年12月25日(周三) 下午5:44 收件人: "SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019"<FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019@noreply.github.com>; 抄送: "Mention"<mention@noreply.github.com>;"baorave"<798976541@qq.com>; 主题: Re: [SoftwareGift/FeatherNets_Face-Anti-spoofing-Attack-Detection-Challenge-CVPR2019] 如何获取实际的预测结果啊？ （how to get predict result for one image?） (#36)

prnet don't fit with embedded device, beacause prnet run only highly with GPU. I choose FeatherNet for run embedded device cause small size trained model. Can you suggest something else for remove background?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

phamthienvuong98 commented 3 years ago

@ApexPredator1 how do you have "model = models.fishnet150()". i dont know. please help me.