1. Which task outputs bounding boxes with labels?
- Semantic segmentation
- Image classification
- Object detection
- OCR only
जवाब देखें
जवाब: Object detection
Detection localises objects with boxes.
Computer Vision कोर्स के 26 सवाल, हर सवाल के सही जवाब और explanation के साथ। जवाब देखने से पहले ख़ुद सोचें।
जवाब: Object detection
Detection localises objects with boxes.
जवाब: (channels, height, width)
Convert with permute when moving from NumPy.
जवाब: It separates colour (hue) from brightness, so masks are more robust to lighting
Hue stays similar when brightness changes.
पाठ: Colour Spaces
जवाब: It reduces noise but also softens edges
Smoothing removes fine detail too.
जवाब: Fewer, stronger edges are kept
Higher thresholds keep only strong gradients.
पाठ: Edge Detection
जवाब: Several touching coins merged into one connected component
Touching objects need separation methods.
जवाब: Fits a geometric transform while rejecting outlier matches
Robust fitting ignores wrong matches.
जवाब: They must be transformed with the same rotation
Labels must follow the pixels.
पाठ: Geometric Transforms
जवाब: So evaluation scores are stable and comparable
Random changes belong in training.
पाठ: Data Augmentation
जवाब: Near-duplicate frames inflate test scores
Keep related images on one side of the split.
जवाब: Residual (skip) connections
Skip connections keep gradients flowing.
जवाब: The model learned from inputs scaled and normalised that way
Mismatched inputs degrade predictions.
जवाब: Only a simple classifier on top of frozen pretrained features
Cheap, fast and often strong.
जवाब: 0.67
Overlap shrinks quickly with shifts.
जवाब: Lower-scoring boxes that heavily overlap a higher-scoring box
It removes duplicates of the same object.
जवाब: Detectors return many low-confidence boxes that are usually wrong
Raw outputs include many weak guesses.
जवाब: Each ground-truth object can be matched only once
Duplicates are penalised.
जवाब: Instance segmentation
Instances are individual objects.
जवाब: Taking the argmax over class scores for each pixel
One class per pixel.
जवाब: Dice is always greater than or equal to IoU
Do not compare one with the other.
जवाब: Confidence can collapse even when the top label stays the same
Falling confidence is an early warning.
जवाब: Pixels in BGR channel order
Convert to RGB if your model expects it.
जवाब: Lower latency, offline operation and keeping images on site
Edge trades capacity for latency and privacy.
जवाब: Text descriptions of the classes, not labelled training images
Images are compared with text prompts.
जवाब: Counting on the device and discarding or blurring faces
Collect and keep only what you need.
जवाब: Identical preprocessing in training and serving, verified with golden images
Consistency and realistic testing.