神经网络图像压缩代码

参加第五届人工智能竞赛，选的图像编码赛道（钱多），纯记录下，这神经网络结构打榜分数也不高，我觉得重要的在于找到一种合适于图像压缩任务的结构，训练倒是其次，主办方让完全采用AI的方式去做，我觉得在网络结构的选取上，势必要加入一些自己对图像的专业理解的，只是这种理解不能以传统的方式表现出来。

点击查看代码

import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import transforms
from PIL import Image
import numpy as np
import lpips
from pytorch_msssim import ms_ssim
from torchvision.transforms.functional import normalizeclass ImageCompressor(nn.Module):def __init__(self, compression_ratio=8):super(ImageCompressor, self).__init__()# 编码器 - 减少通道数以提高压缩率# 编码器部分self.encoder = nn.Sequential(nn.Conv2d(3, 64, kernel_size=3, padding=1),nn.ReLU(),nn.MaxPool2d(2, 2),nn.Conv2d(64, 32, kernel_size=3, padding=1),nn.ReLU(),nn.MaxPool2d(2, 2),nn.Conv2d(32, 16, kernel_size=3, padding=1),nn.ReLU(),nn.MaxPool2d(2, 2),nn.Conv2d(16, 4, kernel_size=3, padding=1),nn.ReLU())# 解码器部分相应修改self.decoder = nn.Sequential(nn.ConvTranspose2d(4, 16, kernel_size=3, padding=1),nn.ReLU(),nn.Upsample(scale_factor=2),nn.ConvTranspose2d(16, 32, kernel_size=3, padding=1),nn.ReLU(),nn.Upsample(scale_factor=2),nn.ConvTranspose2d(32, 64, kernel_size=3, padding=1),nn.ReLU(),nn.Upsample(scale_factor=2),nn.ConvTranspose2d(64, 3, kernel_size=3, padding=1),nn.Sigmoid())def forward(self, x):encoded = self.encoder(x)decoded = self.decoder(encoded)return decodeddef compress_image(image_path, output_path, model_path=None):device = torch.device("cuda" if torch.cuda.is_available() else "cpu")model = ImageCompressor().to(device)# 定义多个损失函数mse_criterion = nn.MSELoss()  # MSE损失（用于PSNR）lpips_criterion = lpips.LPIPS(net='alex').to(device)  # LPIPS损失optimizer = optim.Adam(model.parameters(), lr=0.001)# 准备数据transform = transforms.Compose([transforms.ToTensor(),# 添加归一化处理transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5])])img = Image.open(image_path)img_tensor = transform(img).unsqueeze(0).to(device)# 训练模型model.train()for epoch in range(300):optimizer.zero_grad()output = model(img_tensor)# 计算多个损失# 1. MSE损失（用于PSNR）mse_loss = mse_criterion(output, img_tensor)psnr = -10 * torch.log10(mse_loss)# 2. MS-SSIM损失ms_ssim_loss = 1 - ms_ssim(output, img_tensor, data_range=1.0)# 3. LPIPS损失lpips_loss = lpips_criterion(output, img_tensor).mean()# 组合损失，使用权重平衡各项total_loss = (1.0 * mse_loss +        # 基础重建损失0.3 * ms_ssim_loss +    # 结构相似性损失0.1 * lpips_loss        # 感知损失)total_loss.backward()optimizer.step()if (epoch + 1) % 10 == 0:print(f'Epoch [{epoch+1}/300]')print(f'PSNR: {psnr.item():.2f}')print(f'MS-SSIM Loss: {ms_ssim_loss.item():.4f}')print(f'LPIPS: {lpips_loss.item():.4f}')print('------------------------')# 保存模型if model_path:torch.save(model.state_dict(), model_path)# 压缩图像model.eval()with torch.no_grad():img = Image.open(image_path)img_tensor = transforms.ToTensor()(img).unsqueeze(0).to(device)compressed = model(img_tensor)# 将结果转换回图像output_img = transforms.ToPILImage()(compressed.squeeze(0).cpu())output_img.save(output_path)def decompress_image(compressed_path, output_path, model_path):# 加载模型device = torch.device("cuda" if torch.cuda.is_available() else "cpu")model = ImageCompressor().to(device)if not os.path.exists(model_path):raise ValueError("未找到模型文件！")model.load_state_dict(torch.load(model_path))model.eval()# 解压缩图像with torch.no_grad():img = Image.open(compressed_path)img_tensor = transforms.ToTensor()(img).unsqueeze(0).to(device)decompressed = model(img_tensor)# 将结果转换回图像output_img = transforms.ToPILImage()(decompressed.squeeze(0).cpu())output_img.save(output_path)if __name__ == "__main__":import os# 示例使用input_path = "input.jpg"compressed_path = "compressed.jpg"decompressed_path = "decompressed.jpg"model_path = "compressor_model.pth"# 压缩流程print("正在压缩图像...")compress_image(input_path, compressed_path, model_path)print(f"压缩完成，已保存至 {compressed_path}")# 解压缩流程print("正在解压缩图像...")decompress_image(compressed_path, decompressed_path, model_path)print(f"解压缩完成，已保存至 {decompressed_path}")

目前效果如下：

点击查看代码

Epoch [300/300]
PSNR: 10.01
MS-SSIM Loss: 0.3211
LPIPS: 0.2903