Related Computer Vision Links
Learn Semantic Computer Vision Tutorial, validate concepts with Semantic Computer Vision MCQ Questions, and prepare interviews through Semantic Computer Vision Interview Questions and Answers.
Semantic Segmentation MCQ
Dense labeling: each pixel gets a class. Encoder–decoder architectures, multi-scale context, skip connections, and how metrics differ from detection.
Per-pixel
Class map
Encoder
Downsample
Decoder
Upsample
IoU
Overlap
Semantic segmentation
Every pixel is classified (sky, road, person, …) without separating object instances. Fully convolutional networks replaced sliding-window classifiers for dense outputs.
U-Net idea
Skip connections fuse fine spatial detail from the encoder with semantic features from the bottleneck—sharp boundaries.
Building blocks
FCN
Replace FC layers with 1×1 convs; upsample feature maps to input resolution.
ASPP / dilation
Multi-scale context without losing resolution too aggressively.
Loss
Per-pixel cross-entropy; class imbalance may need weighting or focal-style ideas.
vs instance
Semantic does not separate two people as different IDs—instance segmentation adds that.
Output
H×W label map (or H×W×C logits)