Week 1 ungraded lab - after uploading images, no boxes/object detection

[Xinkai_Zhou]

After trying on both notebook and Docker, I would get to the server page, upload a picture, and the server would return 200, but no object detection – only the original picture without a bounded box or a label. Does anyone know why this is?

[chris.favila]

Hi, and welcome to the Forum! To get more context, can you post a sample result and also mention which exact step in the lab you’re getting this? Did you try using other images, or varying the confidence level? Thanks.

[Xinkai_Zhou]

Hi, thanks! Where do you change the confidence level?

It does say “no parameters” on the page. Perhaps this is the problem? After changing the image, I did get two bounded boxes. But they are redundant.

[chris.favila]

Hi. You can try passing a confidence value in this line before starting the app:

    # 3. RUN OBJECT DETECTION MODEL
    
    # Run object detection
    bbox, label, conf = cv.detect_common_objects(image, model=model, confidence=0.3)

You can also try selecting the yolov3 model instead of the tiny variant and see what results you get.

I’m able to get a box here with confidence=0.1.

Hope this helps.

[sanjaypsachdev]

@Xinkai_Zhou please do let us know if you were able to get the object detection to work in week 1 lab, or if you require further assistance ? Thanks.