Keep aspect for pose resnet

[kyakuno]

pose resnetの入力のアスペクトを保持するようにする。

torch版のオフィシャル実装 https://github.com/microsoft/human-pose-estimation.pytorch https://github.com/microsoft/human-pose-estimation.pytorch/blob/master/lib/dataset/coco.py

    def _xywh2cs(self, x, y, w, h):
        center = np.zeros((2), dtype=np.float32)
        center[0] = x + w * 0.5
        center[1] = y + h * 0.5

        if w > self.aspect_ratio * h:
            h = w * 1.0 / self.aspect_ratio
        elif w < self.aspect_ratio * h:
            w = h * self.aspect_ratio
        scale = np.array(
            [w * 1.0 / self.pixel_std, h * 1.0 / self.pixel_std],
            dtype=np.float32)
        if center[0] != -1:
            scale = scale * 1.25

        return center, scale

tensorflow版のオフィシャル実装 https://github.com/mks0601/TF-SimpleHumanPose https://github.com/mks0601/TF-SimpleHumanPose/blob/master/main/gen_batch.py

    x, y, w, h = bbox
    aspect_ratio = cfg.input_shape[1]/cfg.input_shape[0]
    center = np.array([x + w * 0.5, y + h * 0.5])
    if w > aspect_ratio * h:
        h = w / aspect_ratio
    elif w < aspect_ratio * h:
        w = h * aspect_ratio
    scale = np.array([w,h]) * 1.25
    rotation = 0

エクスポートのIssue https://github.com/axinc-ai/ailia-models/issues/177