1. はじめに

torchvision ライブラリを使用して、ターゲット検出とセマンティックセグメンテーションを実装します。

2. コード

1. ターゲットの検出

from PIL import Image
import matplotlib.pyplot as plt
import torchvision.transforms as T
import torchvision
import numpy as np
import cv2
import random


COCO_INSTANCE_CATEGORY_NAMES = [
    '__background__', 'person', 'bicycle', 'car', 'motorcycle', 'airplane', 'bus',
    'train', 'truck', 'boat', 'traffic light', 'fire hydrant', 'N/A', 'stop sign',
    'parking meter', 'bench', 'bird', 'cat', 'dog', 'horse', 'sheep', 'cow',
    'elephant', 'bear', 'zebra', 'giraffe', 'N/A', 'backpack', 'umbrella', 'N/A', 'N/A',
    'handbag', 'tie', 'suitcase', 'frisbee', 'skis', 'snowboard', 'sports ball',
    'kite', 'baseball bat', 'baseball glove', 'skateboard', 'surfboard', 'tennis racket',
    'bottle', 'N/A', 'wine glass', 'cup', 'fork', 'knife', 'spoon', 'bowl',
    'banana', 'apple', 'sandwich', 'orange', 'broccoli', 'carrot', 'hot dog', 'pizza',
    'donut', 'cake', 'chair', 'couch', 'potted plant', 'bed', 'N/A', 'dining table',
    'N/A', 'N/A', 'toilet', 'N/A', 'tv', 'laptop', 'mouse', 'remote', 'keyboard', 'cell phone',
    'microwave', 'oven', 'toaster', 'sink', 'refrigerator', 'N/A', 'book',
    'clock', 'vase', 'scissors', 'teddy bear', 'hair drier', 'toothbrush'
]


def get_prediction(img_path, threshold):
    # 加载 mask_r_cnn 模型进行目标检测
    model = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True)
    model.eval()
    img = Image.open(img_path)
    transform = T.Compose([T.ToTensor()])
    img = transform(img)
    pred = model([img])
    pred_score = list(pred[0]['scores'].detach().numpy())
    print(pred[0].keys())  # ['boxes', 'labels', 'scores', 'masks']
    pred_t = [pred_score.index(x) for x in pred_score if x > threshold][-1]  # num of boxes
    pred_masks = (pred[0]['masks'] > 0.5).squeeze().detach().cpu().numpy()
    pred_boxes = [[(int(i[0]), int(i[1])), (int(i[2]), int(i[3]))] for i in list(pred[0]['boxes'].detach().numpy())]
    pred_class = [COCO_INSTANCE_CATEGORY_NAMES[i] for i in list(pred[0]['labels'].numpy())]
    pred_masks = pred_masks[:pred_t + 1]
    pred_boxes = pred_boxes[:pred_t + 1]
    pred_class = pred_class[:pred_t + 1]
    return pred_masks, pred_boxes, pred_class


def random_colour_masks(image):
    colours = [[0, 255, 0], [0, 0, 255], [255, 0, 0], [0, 255, 255], [255, 255, 0], [255, 0, 255], [80, 70, 180],
               [250, 80, 190], [245, 145, 50], [70, 150, 250], [50, 190, 190]]
    r = np.zeros_like(image).astype(np.uint8)
    g = np.zeros_like(image).astype(np.uint8)
    b = np.zeros_like(image).astype(np.uint8)
    r[image == 1], g[image == 1], b[image == 1] = colours[random.randrange(0, 10)]
    coloured_mask = np.stack([r, g, b], axis=2)
    return coloured_mask


def instance_segmentation_api(img_path, threshold=0.5, rect_th=3, text_size=2, text_th=2):
    masks, boxes, cls = get_prediction(img_path, threshold)
    img = cv2.imread(img_path)
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    for i in range(len(masks)):
        rgb_mask = random_colour_masks(masks[i])
        randcol = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255))
        img = cv2.addWeighted(img, 1, rgb_mask, 0.5, 0)
        cv2.rectangle(img, boxes[i][0], boxes[i][1], color=randcol, thickness=rect_th)
        cv2.putText(img, cls[i], boxes[i][0], cv2.FONT_HERSHEY_SIMPLEX, text_size, randcol, thickness=text_th)
    plt.figure(figsize=(20, 30))
    plt.imshow(img)
    plt.xticks([])
    plt.yticks([])
    plt.show()
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    cv2.imwrite('result_det.jpg', img)


if __name__ == '__main__':
    instance_segmentation_api('horse.jpg')

2. セマンティックセグメンテーション

import torch
import matplotlib.pyplot as plt
from PIL import Image
import numpy as np
from torchvision import models
from torchvision import transforms


def pre_img(img):
    if img.mode == 'RGBA':
        a = np.asarray(img)[:, :, :3]
        img = Image.fromarray(a)
    return img


def decode_seg_map(image, nc=21):
    label_colors = np.array([(0, 0, 0), (128, 0, 0), (0, 128, 0), (128, 128, 0),
                             (0, 0, 128), (128, 0, 128), (0, 128, 128), (128, 128, 128),
                             (64, 0, 0), (192, 0, 0), (64, 128, 0), (192, 128, 0),
                             (64, 0, 128), (192, 0, 128), (64, 128, 128), (192, 128, 128),
                             (0, 64, 0), (128, 64, 0), (0, 192, 0), (128, 192, 0), (0, 64, 128)])
    r = np.zeros_like(image).astype(np.uint8)
    g = np.zeros_like(image).astype(np.uint8)
    b = np.zeros_like(image).astype(np.uint8)

    for l in range(0, nc):
        idx = image == l
        r[idx] = label_colors[l, 0]
        g[idx] = label_colors[l, 1]
        b[idx] = label_colors[l, 2]

    return np.stack([r, g, b], axis=2)


if __name__ == '__main__':
    # 加载 deep_lab_v3 模型进行语义分割
    model = models.segmentation.deeplabv3_resnet101(pretrained=True)
    model = model.eval()

    img = Image.open('horse.jpg')
    print(img.size)  # (694, 922)
    plt.imshow(img)
    plt.axis('off')
    plt.show()

    im = pre_img(img)
    transform = transforms.Compose([
        transforms.Resize(256),
        transforms.CenterCrop(224),
        transforms.ToTensor(),
        transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
    ])
    input_img = transform(im).unsqueeze(0)  # resize
    tt = np.transpose(input_img.detach().numpy()[0], (1, 2, 0))  # transpose
    print(tt.shape)  # (224, 224, 3)
    plt.imshow(tt)
    plt.axis('off')
    plt.show()

    output = model(input_img)
    print(output.keys())  # odict_keys(['out', 'aux'])
    print(output['out'].shape)  # torch.Size([1, 21, 224, 224])
    output = torch.argmax(output['out'].squeeze(), dim=0).detach().cpu().numpy()
    result_class = set(list(output.flat))
    print(result_class)  # {0, 13, 15}

    rgb = decode_seg_map(output)
    print(rgb.shape)  # (224, 224, 3)
    img = Image.fromarray(rgb)
    img.save('result_seg.jpg')
    plt.axis('off')
    plt.imshow(img)
    plt.show()