293 - Denoising RGB images using deep learning (Noise2Void)

Поділитися
Вставка
  • Опубліковано 28 жов 2024

КОМЕНТАРІ • 21

  • @jurischaber6935
    @jurischaber6935 Рік тому +1

    Der DigitalSreeni,
    I am a radioligist and yesterday I passed my Master thesis in medical computer science. Your Videos really helped me a lot.
    Thanks a lot for helping me to become Master of Science.

    • @DigitalSreeni
      @DigitalSreeni  Рік тому

      That is wonderful, congratulations!!! I am glad you found my tutorials to be useful towards your Masters work.

  • @talhazia1401
    @talhazia1401 3 місяці тому +1

    That is NOT a halo but a real object. Its called Helix Nebula that is a planetary nebula meaning a star once shone is now dead at its center and leftover material is what you see in red and green colors.

  • @cplusplus-python
    @cplusplus-python Рік тому +1

    Fantastic job, Thanks, Prof. Noise2Void seems to be an excellent method to remove noises also... I might be more curious about how to balance the colours using our model, for example, white balancing, contrast or gamma?

    • @DigitalSreeni
      @DigitalSreeni  Рік тому +1

      Various approaches, not sure if there is a library that does it automatically but it should be easy to code. For example, please look up the grayworld approximation technique for white balancing. It just involves getting the average color and comparing it to gray.

  • @bijaynayak6473
    @bijaynayak6473 11 місяців тому

    Hello Sreeni ,big fan of your teaching channel, can u create any model which will help to create image orientation and perspectives correction for ocr based images

  • @TheLegend706
    @TheLegend706 Рік тому

    Hey! following your u-net semantic segmentation videos, and i found a video about how to limit my gpu usage while using tensorflow bur i guess it's outdated and i keep getting errors, can you help me?

  • @chitti9974
    @chitti9974 Рік тому

    Sir, iam a research scholar..I worked and executed code you provided to train Srgan...by 32x32 lr images and 128x128 hr images..with 50 epochs.using the saved model I tried to get hr images for 224x224..but it gives error and displays that input is only 32x32. How to generalise the model sir???. If we give any size input it should give output image 4 times of the input. Sir pls reply....super resolution is my phd topic...and..my execution stopped at this point

  • @user-maomao-tsai
    @user-maomao-tsai Рік тому

    Dear DigitalSreeni , may I ask a question about "SyntaxError: unexpected EOF while parsing" It happened on 214 , 227 for ...for , for .... if , Thanks

  • @fikretburhanylmaz6422
    @fikretburhanylmaz6422 Місяць тому

    Teşekkürler.

  • @meriemchahinez9373
    @meriemchahinez9373 Рік тому

    Hi! Can you please make a tutorial about concatenation of two CNN models with a different inputs

  • @buffetbar8231
    @buffetbar8231 Рік тому

    How to decide the number of hidden layers and nodes in a hidden layer?

  • @sm-op9zg
    @sm-op9zg Рік тому

    sir.. please make a video about SOTA architecture.. please🙏

  • @KrishenJagani
    @KrishenJagani Рік тому

    Hi! can you make a tutorial on saliency prediction?

  • @samadhansurvase4804
    @samadhansurvase4804 Рік тому

    Hello hii i am working on project like roomvo , visualise the room by fill tiles in floor ,wall, so for that here I used canny to detect the edges but canny doesn't work properly so what should I do pls give me sol
    And instead of canny is any

    • @DigitalSreeni
      @DigitalSreeni  Рік тому

      You can try this: ua-cam.com/video/UIrvEG9Oj1s/v-deo.html

  • @mardeenosman8979
    @mardeenosman8979 Рік тому

    Thanks alot, What about denoising DICOM images??

    • @DigitalSreeni
      @DigitalSreeni  Рік тому

      This approach will work to denoise DICOM images, as long as the noise is unstructured.

  • @ernestbeckham2921
    @ernestbeckham2921 Рік тому

    hello sir,
    your videos are helpful to me. can you upload video about skull stripping?

  • @mohammedalkhowaiter7562
    @mohammedalkhowaiter7562 Рік тому

    Thank you for all your videos! Its' a school for whole the world! I have a question if you see my comments. Why every time I export an image that is been transformed from RGB into a tensor pytorch, to save it back from the tensor to RGB again and test it, it gives different class sometimes? I use torchvision.utils.save_image library. Thanks in advance!