I used the following code to decode the image in png format into the allocated memory block *imageData. - (void)decodeImage:(UIImage*)image { GLubyte* imageData = (GLubyte*)malloc(image.size.width * image.size.height * 4); CGColorSpaceRef colorSpace = CGColorSpaceCreateDeviceRGB(); CGContextRef imageContext = CGBitmapContextCreate(imageData, image.size.width, image.size.height, 8, image.size.width * 4, colorSpace, kCGImageAlphaPremultipliedLast | kCGBitmapByteOrderDefault); CGContextDrawImage(imageContext, CGRectMake(0.0, 0.0, image.size.width, image.size.height), image.CGImage); CGContextRelease(imageContext); CGColorSpaceRelease(colorSpace); int bytesPerRow = image.size.width * 4; //you can log [330, 150] RGBA value here, then gets wrong alpha value; int targetRow = 330; int targetCol = 150; u_int32_t r = (u_int32_t)imageData[targetRow*bytesPerRow*4 + targetCol*4 + 0]; u_int32_t g = (u_int32_t)imageData[targetRow*bytesPerRow*4 + targetCol*4 + 1]; u_int32_t b = (u_int32_t)imageData[targetRow*bytesPerRow*4 + targetCol*4 + 2]; u_int32_t a = (u_int32_t)imageData[targetRow*bytesPerRow*4 + targetCol*4 + 3]; free(imageData); } The CGBitmapInfo used is kCGImageAlphaPremultipliedLast | kCGBitmapByteOrderDefault. The problem is that the alpha channel is lost in the decoded result: For example, the RGBA values ​​of the target png at [row, col] = [330, 150] are R = 240, B = 125, G = 106, A = 80. If the decoding is correct, the expected result should be R = 75, G = 39, B = 33, A = 80 with AlphaPremultiplied. However, after decoding the png on iOS17, the result is R = 75, G = 39, B = 33, A = 255, where the alpha values ​​are all forced to 255. Xcode Version 15.0 beta (15A5160n). iPhone14 Pro. The png file: