Malaysian Trash Images with Instance Segmentation Annotations

Annotated images of trash in Malaysia, using instance segmentation to fit real-world noise. The dataset is hosted on Kaggle, but the author, organization, and specific collection details are not provided. The temporal coverage, total number of images, and file formats are unknown.

Use Cases

• Train instance segmentation models to identify and delineate individual pieces of trash.
• Benchmark model robustness against real-world noise and occlusion in waste imagery.
• Develop automated waste sorting or recycling systems based on visual recognition.

Strengths

• Annotations use instance segmentation, which provides pixel-level masks for individual objects.
• Data is described as fitting real-world noise, suggesting practical relevance for model deployment.

Limitations

• Row count and dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Kaggle

Collection Method

Likely contains images annotated with instance segmentation masks.

Geography

Malaysia