png图像生成

[andre20000131]

代码里也没有说最后咋生成的jpg/png啊？只有一个点云pccc的训练，有很多细节都没有说清。。。

[xyxingx]

Hi,
The goal of PCCC is to estimate the illumination from the input point cloud. Therefore, we only output the predicted illumination in the training script. You can visualize the white-balanced image by simply dividing the raw input image with the estimated illumination. One example code shows as follows:
img_awb,pred_pic = get_awb_pic(img_raw,pred) pred_pic = np.clip(pred_pic, 0, 1) img_awb_single,pred_weights,weights = get_awb_pic_single(img_raw,pred,pred_pic)
where image_raw is the raw image (might need to resize), pred is estimated by PCCC. get_awb_pic outputs the per-pixel illumination map (pred_pic) and the recovered image (img_awb). get_awb_pic_single outputs the image recovered by the single illumination (img_awb_single).
note: all images are gamma-corrected.

[andre20000131]

thank you soooo much, i finished all process.

[andre20000131]

i am very interesting about your work, i want to make more communication with you,can i get your email...thankyou

[xyxingx]

sure, my email is x.xing at uva dot nl Fuchen ***@***.***> 于2024年2月29日周四 07:48写道：

…

i am very interesting about your work, i want to make more communication with you,can i get your email...thankyou — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AMQ2OHLFGPGSS6MCC62SGKLYV3HLXAVCNFSM6AAAAABD5OEEPKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSNZQGUYTEMRUGQ> . You are receiving this because you commented.Message ID: ***@***.***>