Hello author, I am getting an error when running main_test, py file

[Feiyuanyu]

Hello, I have reported an error while running the main_test.py file without changing any data or code.Hello, I have reported an error while running the main_test.py file without changing any data or code.

[Feiyuanyu]

I only get two images in the results folder, and the two images in the result folder are always changing when I run the main_test.py file.

[DeepTrial]

sorry to be late. i thought u have to check that whether the main_test.py code have load all the test images and their groundtruths. i do the test in win10 and maybe there are some difference on ubuntu.

[fubai3]

您好，我用自己的图片测试的时候遇到了同样的问题，请问您是怎么解决的

[fubai3]

您好，我用自己的图片测试的时候遇到了同样的问题，请问您是怎么解决的

[SIAT-liyang]

I only get two images in the results folder, and the two images in the result folder are always changing when I run the main_test.py file.

I have the same problem. Because of the path, there is a problem with the code here in XXX. The file name is not read out. All the saved files will be overwritten by the saved files. Here is my solution(in segmention_infer.py): ` def predict(self):

    predList=glob.glob(self.config.test_img_path+"*."+self.config.test_datatype)

    i = 0

    for path in predList:

        orgImg_temp=plt.imread(path)
        orgImg=orgImg_temp[:,:,1]*0.75+orgImg_temp[:,:,0]*0.25
        print("[Info] Analyze filename...",self.analyze_name(path))
        height,width=orgImg.shape[:2]
        orgImg = np.reshape(orgImg, (height,width,1))
        patches_pred,new_height,new_width,adjustImg=get_test_patches(orgImg,self.config)
        predictions = self.model.predict(patches_pred, batch_size=32, verbose=1)
        pred_patches=pred_to_patches(predictions,self.config)
        pred_imgs=recompone_overlap(pred_patches,self.config,new_height,new_width)
        pred_imgs=pred_imgs[:,0:height,0:width,:]
        adjustImg=adjustImg[0,0:height,0:width,:]
        print(adjustImg.shape)
        probResult=pred_imgs[0,:,:,0]
        binaryResult=gray2binary(probResult)
        resultMerge=visualize([adjustImg,binaryResult],[1,2])
        resultMerge=cv2.cvtColor(resultMerge,cv2.COLOR_RGB2BGR)
        i=i+1
        cv2.imwrite(self.config.test_result_path+str(i)+"_merge.jpg",resultMerge)
        cv2.imwrite(self.config.test_result_path+str(i)+ "_prob.bmp", (probResult*255).astype(np.uint8))

`

[fubai3]

谢谢您的回复，您训练的时候loss能下降到多少，我有110张训练图片，65张验证图片，训练了120个epoch后结果如下

测试效果并不好，请问是epoch少了还是什么原因呢 ------------------ 原始邮件 ------------------ 发件人: "SIAT-liyang"notifications@github.com; 发送时间: 2019年10月28日(星期一) 下午4:39 收件人: "DeepTrial/Retina-VesselNet"Retina-VesselNet@noreply.github.com; 抄送: "℡ 深海的鱼"1165269652@qq.com;"Comment"comment@noreply.github.com; 主题: Re: [DeepTrial/Retina-VesselNet] Hello author, I am getting an errorwhen running main_test, py file (#33)

Hello, I have reported an error while running the main_test.py file without changing any data or code.Hello, I have reported an error while running the main_test.py file without changing any data or code.

I have the same problem. Because of the path, there is a problem with the code here in XXX. The file name is not read out. All the saved files will be overwritten by the saved files. Here is my solution(in segmention_infer.py)： def predict(self): predList=glob.glob(self.config.test_img_path+"*."+self.config.test_datatype) i = 0 for path in predList: orgImg_temp=plt.imread(path) orgImg=orgImg_temp[:,:,1]*0.75+orgImg_temp[:,:,0]*0.25 print("[Info] Analyze filename...",self.analyze_name(path)) height,width=orgImg.shape[:2] orgImg = np.reshape(orgImg, (height,width,1)) patches_pred,new_height,new_width,adjustImg=get_test_patches(orgImg,self.config) predictions = self.model.predict(patches_pred, batch_size=32, verbose=1) pred_patches=pred_to_patches(predictions,self.config) pred_imgs=recompone_overlap(pred_patches,self.config,new_height,new_width) pred_imgs=pred_imgs[:,0:height,0:width,:] adjustImg=adjustImg[0,0:height,0:width,:] print(adjustImg.shape) probResult=pred_imgs[0,:,:,0] binaryResult=gray2binary(probResult) resultMerge=visualize([adjustImg,binaryResult],[1,2]) resultMerge=cv2.cvtColor(resultMerge,cv2.COLOR_RGB2BGR) **i=i+1** cv2.imwrite(self.config.test_result_path+**str(i)**+"_merge.jpg",resultMerge) cv2.imwrite(self.config.test_result_path+**str(i)**+ "_prob.bmp", (probResult*255).astype(np.uint8))

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

[SIAT-liyang]

您好，我用自己的图片测试的时候遇到了同样的问题，请问您是怎么解决的

我也遇到了同样的问题。由于路径的原因，segmention_infer.py最后两行的代码有问题，没有读取出文件名，所有每次保存的文件将原来已经保存的文件覆盖掉了，添加一个计数器i来重新命名就行了

[SIAT-liyang]

谢谢您的回复，您训练的时候loss能下降到多少，我有110张训练图片，65张验证图片，训练了120个epoch后结果如下 测试效果并不好，请问是epoch少了还是什么原因呢 ------------------ 原始邮件 ------------------ 发件人: "SIAT-liyang"notifications@github.com; 发送时间: 2019年10月28日(星期一) 下午4:39 收件人: "DeepTrial/Retina-VesselNet"Retina-VesselNet@noreply.github.com; 抄送: "℡ 深海的鱼"1165269652@qq.com;"Comment"comment@noreply.github.com; 主题: Re: [DeepTrial/Retina-VesselNet] Hello author, I am getting an errorwhen running main_test, py file (#33) Hello, I have reported an error while running the main_test.py file without changing any data or code.Hello, I have reported an error while running the main_test.py file without changing any data or code. I have the same problem. Because of the path, there is a problem with the code here in XXX. The file name is not read out. All the saved files will be overwritten by the saved files. Here is my solution(in segmention_infer.py)： def predict(self): predList=glob.glob(self.config.test_img_path+"*."+self.config.test_datatype) i = 0 for path in predList: orgImg_temp=plt.imread(path) orgImg=orgImg_temp[:,:,1]*0.75+orgImg_temp[:,:,0]*0.25 print("[Info] Analyze filename...",self.analyze_name(path)) height,width=orgImg.shape[:2] orgImg = np.reshape(orgImg, (height,width,1)) patches_pred,new_height,new_width,adjustImg=get_test_patches(orgImg,self.config) predictions = self.model.predict(patches_pred, batch_size=32, verbose=1) pred_patches=pred_to_patches(predictions,self.config) pred_imgs=recompone_overlap(pred_patches,self.config,new_height,new_width) pred_imgs=pred_imgs[:,0:height,0:width,:] adjustImg=adjustImg[0,0:height,0:width,:] print(adjustImg.shape) probResult=pred_imgs[0,:,:,0] binaryResult=gray2binary(probResult) resultMerge=visualize([adjustImg,binaryResult],[1,2]) resultMerge=cv2.cvtColor(resultMerge,cv2.COLOR_RGB2BGR) **i=i+1** cv2.imwrite(self.config.test_result_path+**str(i)**+"_merge.jpg",resultMerge) cv2.imwrite(self.config.test_result_path+**str(i)**+ "_prob.bmp", (probResult*255).astype(np.uint8)) — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

看不到你的测试效果

[fubai3]

您好，这是我训练了120个epoch后的日志和测试图片，效果并不好，是epoch少了还是其他原因呢，谢谢 ------------------ 原始邮件 ------------------ 发件人: "SIAT-liyang"notifications@github.com; 发送时间: 2019年10月28日(星期一) 下午4:51 收件人: "DeepTrial/Retina-VesselNet"Retina-VesselNet@noreply.github.com; 抄送: "℡ 深海的鱼"1165269652@qq.com;"Comment"comment@noreply.github.com; 主题: Re: [DeepTrial/Retina-VesselNet] Hello author, I am getting an errorwhen running main_test, py file (#33)

谢谢您的回复，您训练的时候loss能下降到多少，我有110张训练图片，65张验证图片，训练了120个epoch后结果如下 测试效果并不好，请问是epoch少了还是什么原因呢 ------------------ 原始邮件 ------------------ 发件人: "SIAT-liyang"notifications@github.com; 发送时间: 2019年10月28日(星期一) 下午4:39 收件人: "DeepTrial/Retina-VesselNet"Retina-VesselNet@noreply.github.com; 抄送: "℡ 深海的鱼"1165269652@qq.com;"Comment"comment@noreply.github.com; 主题: Re: [DeepTrial/Retina-VesselNet] Hello author, I am getting an errorwhen running main_test, py file (#33) Hello, I have reported an error while running the main_test.py file without changing any data or code.Hello, I have reported an error while running the main_test.py file without changing any data or code. I have the same problem. Because of the path, there is a problem with the code here in XXX. The file name is not read out. All the saved files will be overwritten by the saved files. Here is my solution(in segmention_infer.py)： def predict(self): predList=glob.glob(self.config.test_img_path+"."+self.config.test_datatype) i = 0 for path in predList: orgImg_temp=plt.imread(path) orgImg=orgImg_temp[:,:,1]0.75+orgImg_temp[:,:,0]0.25 print("[Info] Analyze filename...",self.analyze_name(path)) height,width=orgImg.shape[:2] orgImg = np.reshape(orgImg, (height,width,1)) patches_pred,new_height,new_width,adjustImg=get_test_patches(orgImg,self.config) predictions = self.model.predict(patches_pred, batch_size=32, verbose=1) pred_patches=pred_to_patches(predictions,self.config) pred_imgs=recompone_overlap(pred_patches,self.config,new_height,new_width) pred_imgs=pred_imgs[:,0:height,0:width,:] adjustImg=adjustImg[0,0:height,0:width,:] print(adjustImg.shape) probResult=pred_imgs[0,:,:,0] binaryResult=gray2binary(probResult) resultMerge=visualize([adjustImg,binaryResult],[1,2]) resultMerge=cv2.cvtColor(resultMerge,cv2.COLOR_RGB2BGR) i=i+1 cv2.imwrite(self.config.test_result_path+str(i)+"_merge.jpg",resultMerge) cv2.imwrite(self.config.test_result_path+str(i)+ "_prob.bmp", (probResult255).astype(np.uint8)) — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

看不到你的测试效果

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

[fubai3]

您好，方便加您qq或微信咨询吗，我发了图片没显示

------------------ 原始邮件 ------------------ 发件人: "SIAT-liyang"notifications@github.com; 发送时间: 2019年10月28日(星期一) 下午4:51 收件人: "DeepTrial/Retina-VesselNet"Retina-VesselNet@noreply.github.com; 抄送: "℡ 深海的鱼"1165269652@qq.com;"Comment"comment@noreply.github.com; 主题: Re: [DeepTrial/Retina-VesselNet] Hello author, I am getting an errorwhen running main_test, py file (#33)

谢谢您的回复，您训练的时候loss能下降到多少，我有110张训练图片，65张验证图片，训练了120个epoch后结果如下 测试效果并不好，请问是epoch少了还是什么原因呢 ------------------ 原始邮件 ------------------ 发件人: "SIAT-liyang"notifications@github.com; 发送时间: 2019年10月28日(星期一) 下午4:39 收件人: "DeepTrial/Retina-VesselNet"Retina-VesselNet@noreply.github.com; 抄送: "℡ 深海的鱼"1165269652@qq.com;"Comment"comment@noreply.github.com; 主题: Re: [DeepTrial/Retina-VesselNet] Hello author, I am getting an errorwhen running main_test, py file (#33) Hello, I have reported an error while running the main_test.py file without changing any data or code.Hello, I have reported an error while running the main_test.py file without changing any data or code. I have the same problem. Because of the path, there is a problem with the code here in XXX. The file name is not read out. All the saved files will be overwritten by the saved files. Here is my solution(in segmention_infer.py)： def predict(self): predList=glob.glob(self.config.test_img_path+"."+self.config.test_datatype) i = 0 for path in predList: orgImg_temp=plt.imread(path) orgImg=orgImg_temp[:,:,1]0.75+orgImg_temp[:,:,0]0.25 print("[Info] Analyze filename...",self.analyze_name(path)) height,width=orgImg.shape[:2] orgImg = np.reshape(orgImg, (height,width,1)) patches_pred,new_height,new_width,adjustImg=get_test_patches(orgImg,self.config) predictions = self.model.predict(patches_pred, batch_size=32, verbose=1) pred_patches=pred_to_patches(predictions,self.config) pred_imgs=recompone_overlap(pred_patches,self.config,new_height,new_width) pred_imgs=pred_imgs[:,0:height,0:width,:] adjustImg=adjustImg[0,0:height,0:width,:] print(adjustImg.shape) probResult=pred_imgs[0,:,:,0] binaryResult=gray2binary(probResult) resultMerge=visualize([adjustImg,binaryResult],[1,2]) resultMerge=cv2.cvtColor(resultMerge,cv2.COLOR_RGB2BGR) i=i+1 cv2.imwrite(self.config.test_result_path+str(i)+"_merge.jpg",resultMerge) cv2.imwrite(self.config.test_result_path+str(i)+ "_prob.bmp", (probResult255).astype(np.uint8)) — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

看不到你的测试效果

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

[xihuanliuliu]

大佬们，请问这个准确率ACC，F1-score怎么计算啊？他这个TP有255个值，我就不知道该怎么计算了

[xihuanliuliu]

大佬们，请问这个准确率ACC，F1-score怎么计算啊？他这个TP有255个值，我就不知道该怎么计算了

[manvirvirk]

how to train model without using pre trained model?