原創(chuàng)：余曉龍

Pytorch中提供了很多已經(jīng)在ImageNet數(shù)據(jù)集上訓(xùn)練好的模型了，可以直接被加載到模型中進(jìn)行預(yù)測(cè)任務(wù)稳懒。預(yù)訓(xùn)練模型存放在Pytorch的torchvision中庫(kù)闲擦，在torchvision庫(kù)的models模塊下可以查看內(nèi)置的模型，models模塊中的模型包含四大類场梆，如圖所示：

一墅冷、圖像分類代碼實(shí)現(xiàn)

# coding: utf-8

from PIL import Image
import matplotlib.pyplot as plt
plt.rcParams['font.sans-serif'] = ['SimHei']  # 步驟一（替換sans-serif字體）
plt.rcParams['axes.unicode_minus'] = False   # 步驟二（解決坐標(biāo)軸負(fù)數(shù)的負(fù)號(hào)顯示問(wèn)題）
import json
import numpy as np

import torch
import torch.nn.functional as F
from torchvision import models, transforms

# 1.下載并加載預(yù)訓(xùn)練模型
model = models.resnet18(pretrained=True)
model = model.eval()

# 2.加載標(biāo)簽并對(duì)輸入數(shù)據(jù)進(jìn)行處理
labels_path = './imagenet_class_index.json'
with open(labels_path) as json_data:
    idx2labels = json.load(json_data)
# print(idx2labels)


def getone(onestr):
    return onestr.replace(',', ' ')


# 加載中文標(biāo)簽
with open('./zh_label.csv', 'r+', encoding='gbk') as f:
    # print(f)
    # print(map(getone, list(f)))
    zh_labels = list(map(getone, list(f)))

    print(len(zh_labels), type(zh_labels), zh_labels[:5])


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]
    )
])

# 3.使用模型進(jìn)行預(yù)測(cè)
def preimg(img):
    if img.mode == 'RGBA':
        ch = 4
        a = np.asarray(img)[:, :, :3]
        img = Image.fromarray(a)
    return img


im = preimg(Image.open('panda.jpg'))
transformed_img = transform(im)

inputimg = transformed_img.unsqueeze(0)

output = model(inputimg)
output = F.softmax(output, dim=1)

prediction_score, pred_label_idx = torch.topk(output, 3)
prediction_score = prediction_score.detach().numpy()[0]
print(prediction_score[0])

pred_label_idx = pred_label_idx.detach().numpy()[0]
print(pred_label_idx)

predicted_label = idx2labels[str(pred_label_idx[0])][1]
print(predicted_label)

predicted_label_zh = zh_labels[pred_label_idx[0] + 1]
print(predicted_label_zh)

# 4.預(yù)測(cè)結(jié)果可視化
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 8))
fig.sca(ax1)
ax1.imshow(im)
plt.xticks([])
plt.yticks([])

barlist = ax2.bar(range(3), [i for i in prediction_score])
barlist[0].set_color('g')

plt.sca(ax2)
plt.ylim([0, 1.1])

plt.xticks(range(3),
           # [idx2labels[str(i)][1] for i in pred_label_idx],
           [zh_labels[pred_label_idx[i] + 1] for i in range(3)],
           rotation='45')
fig.subplots_adjust(bottom=0.2)
plt.show()

圖像分類結(jié)果圖

輸入一張熊貓圖片，右邊輸出模型的預(yù)測(cè)結(jié)果或油，如上圖所示寞忿。

二、目標(biāo)檢測(cè)代碼實(shí)現(xiàn)

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

# 加載maskrcnn模型進(jìn)行目標(biāo)檢測(cè)
model = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True)
model.eval()
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):
    img = Image.open(img_path)
    transform = T.Compose([T.ToTensor()])
    img = transform(img)
    pred = model([img])
    print('pred')
    print(pred)
    pred_score = list(pred[0]['scores'].detach().numpy())
    pred_t = [pred_score.index(x) for x in pred_score if x > threshold][-1]
    print("masks>0.5")
    print(pred[0]['masks'] > 0.5)
    masks = (pred[0]['masks'] > 0.5).squeeze().detach().cpu().numpy()
    print("this is masks")
    print(masks)
    pred_class = [COCO_INSTANCE_CATEGORY_NAMES[i] for i in list(pred[0]['labels'].numpy())]
    pred_boxes = [[(i[0], i[1]), (i[2], i[3])] for i in list(pred[0]['boxes'].detach().numpy())]
    masks = masks[:pred_t + 1]
    pred_boxes = pred_boxes[:pred_t + 1]
    pred_class = pred_class[:pred_t + 1]
    return 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=10, text_th=3):
    masks, boxes, pred_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, randcol = random_colour_masks(masks[i]), (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, pred_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()


instance_segmentation_api('./horse.jpg')

原始圖像

maskrcnn模型預(yù)測(cè)圖片

三顶岸、語(yǔ)義分割代碼實(shí)現(xiàn)

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

# 加載deeplabv3模型進(jìn)行語(yǔ)義分割
model = models.segmentation.deeplabv3_resnet101(pretrained=True)
model = model.eval()

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]
    )
])

def preimg(img):
    if img.mode == 'RGBA':
        ch = 4
        a = np.asarray(img)[:, :, :3]
        img = Image.fromarray(a)
    return img

img = Image.open('./horse.jpg')
plt.imshow(img)
plt.axis('off')
plt.show()

im = preimg(img)

inputimg = transform(im).unsqueeze(0)

tt = np.transpose(inputimg.detach().numpy()[0], (1, 2, 0))
plt.imshow(tt)
plt.axis('off')
plt.show()


output = model(inputimg)
print(output['out'].shape)


output = torch.argmax(output['out'].squeeze(),
                      dim=0).detach().cpu().numpy()

resultclass = set(list(output.flat))
print(resultclass)


def decode_segmap(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)


rgb = decode_segmap(output)
print(rgb)

img = Image.fromarray(rgb)
plt.axis('off')
plt.imshow(img)
plt.show()

預(yù)處理后的圖片

語(yǔ)義分割后的圖片

模型從圖中識(shí)別出了兩個(gè)類別的內(nèi)容腔彰，索引值分別為13、15辖佣，所對(duì)應(yīng)的分類名稱是馬和人萍桌。調(diào)用函數(shù)，通過(guò)對(duì)預(yù)測(cè)結(jié)果進(jìn)行染色凌简，得到的預(yù)測(cè)結(jié)果如上圖所示上炎。