T2映射后DICOM图像显示异常，求修复以正常显示

需求说明 #

需要实现以下DICOM图像T2映射功能：

• 按TE2 > TE1规则对指定路径下的DCM文件配对
• 通过公式 T2=(TE2-TE1)/(ln(S1)-ln(S2)) 计算每个像素的T2值
• 对所有符合条件的文件对重复计算
• 将每对T2值按像素累加后求平均值
• 将平均T2值重建为8位DICOM图像，在Scene2中完整显示并保存

当前代码问题 #

生成的T2映射图像在Scene2中显示错乱，无法和左侧DICOM图像一样正常展示，核心原因包括：T2值直接转uint8导致数值截断、QImage内存布局解析错误、缺少异常值处理。

修复后的完整代码 #

def T2_Mapping(self):
    folder_path = self.TE_lineedit.text()
    dicom_files = [file_name for file_name in os.listdir(folder_path)
                   if os.path.isfile(os.path.join(folder_path, file_name)) and file_name.endswith('.dcm')]
    dicom_files.sort()
    pixel_t2_values = []  # 改用列表存储，提升计算效率

    for i in range(len(dicom_files) - 1):
        dicom1 = pydicom.dcmread(os.path.join(folder_path, dicom_files[i]))
        te1 = dicom1.EchoTime
        signal1 = dicom1.pixel_array.astype(np.float32)

        for j in range(i + 1, len(dicom_files)):
            dicom2 = pydicom.dcmread(os.path.join(folder_path, dicom_files[j]))
            te2 = dicom2.EchoTime
            if te2 > te1:
                signal2 = dicom2.pixel_array.astype(np.float32)
                # 处理log(0)异常，限制T2值在合理医学范围
                epsilon = 1e-6
                signal1_safe = np.maximum(signal1, epsilon)
                signal2_safe = np.maximum(signal2, epsilon)
                t2_map = (te2 - te1) / (np.log(signal1_safe) - np.log(signal2_safe))
                t2_map = np.clip(t2_map, 0, 1000)
                pixel_t2_values.append(t2_map)
    
    if not pixel_t2_values:
        print("未找到符合TE2>TE1的有效文件对")
        return
    
    # 计算平均T2值
    averaged_t2_array = np.mean(pixel_t2_values, axis=0)
    
    # 将T2值归一化到0-255区间，避免直接转uint8丢失细节
    t2_min = averaged_t2_array.min()
    t2_max = averaged_t2_array.max()
    if t2_max > t2_min:
        normalized_t2 = ((averaged_t2_array - t2_min) / (t2_max - t2_min)) * 255
    else:
        normalized_t2 = np.zeros_like(averaged_t2_array)
    new_pixel_array = normalized_t2.astype(np.uint8)
    
    # 生成标准化的8位DICOM文件
    new_dicom = dicom1.copy()
    new_dicom.PixelData = new_pixel_array.tobytes()
    new_dicom.Rows, new_dicom.Columns = new_pixel_array.shape
    new_dicom.BitsAllocated = 8
    new_dicom.BitsStored = 8
    new_dicom.HighBit = 7
    new_dicom.SamplesPerPixel = 1
    new_dicom.PhotometricInterpretation = "MONOCHROME2"
    new_dicom.PixelRepresentation = 0
    new_dicom.RescaleIntercept = 0
    new_dicom.RescaleSlope = 1
    output_file_path = os.path.join(folder_path, "t2_mapping.dcm")
    new_dicom.save_as(output_file_path)
    
    # 正确读取并显示图像，适配Scene2布局
    new_dicom = pydicom.dcmread(output_file_path)
    pixel_array = new_dicom.pixel_array
    # 指定字节行间距，确保QImage正确解析像素内存布局
    qimage = QImage(pixel_array.data, pixel_array.shape[1], pixel_array.shape[0], 
                    pixel_array.strides[0], QImage.Format_Grayscale8)
    pixmap = QPixmap.fromImage(qimage)
    
    self.scene2.clear()
    pixmap_item = QGraphicsPixmapItem(pixmap)
    self.scene2.addItem(pixmap_item)
    self.scene2.setSceneRect(pixmap_item.boundingRect())
    # 自适应视图大小，和左侧图像显示逻辑保持一致
    self.view2.fitInView(pixmap_item, Qt.KeepAspectRatio)
    self.view2.setScene(self.scene2)

关键修复点 #

1. 数值范围处理：将T2值归一化到0-255区间，避免直接转uint8导致的数值截断
2. 异常值防护：添加epsilon避免log(0)错误，用clip限制T2值在医学合理范围
3. QImage内存解析：指定strides[0]行间距，确保图像数据正确映射
4. 显示逻辑对齐：使用fitInView让视图自适应图像尺寸，和左侧显示逻辑保持一致

内容的提问来源于stack exchange，提问作者SUAH YU