Billy Kamara ART
Studio
BILLY KAMARA VISUAL AI ANNOTATION PORTFOLIO
Visual AI Data Annotator | Image Quality Specialist | Visual Artist & Designer Fine Arts-trained visual professional with 25+ years of experience in visual analysis, image editing, design, technical drawing, and production quality control. This portfolio documents hands-on CVAT practice in object detection, polygon segmentation, precision mask segmentation, occlusion handling, ontology adherence, and annotation QA.
Platform: CVAT
Focus: Visual Annotation & QA
Location: Alexandria, Virginia
billykamara1@hotmail.com | billykamaraart.com
Phone: +1 (860) 249-4317
This is your About page. This space is a great opportunity to give a full background on who you are, what you do and what your site has to offer. Your users are genuinely interested in learning more about you, so don’t be afraid to share personal anecdotes to create a more friendly quality. Every website has a story, and your visitors want to hear yours. This space is a great opportunity to provide any personal details you want to share with your followers. Include interesting anecdotes and facts to keep readers engaged. Double click on the text box to start editing your content and make sure to add all the relevant details you want site visitors to know. If you’re a business, talk about how you started and share your professional journey. Explain your core values, your commitment to customers and how you stand out from the crowd. Add a photo, gallery or video for even more engagement.
01 Interior Object Detection
Method: Bounding-box annotation Tool: CVAT
Project objective
Identify and localize visible instances across six defined classes: Painting, Chair, Table, Light Fixture, Sofa, and Decorative Object. Bounding boxes were kept tight to visible object extent, with hidden geometry excluded rather than inferred.
Results & QA
21 physical object instances identified and 28 annotation regions created. A second QA pass produced two occlusion-related boundary corrections. One lamp classification was flagged as ambiguous rather than forced into an uncertain class.
Demonstrated skills
Multi-class object detection, tight bounding boxes, instance separation, occlusion handling, visible-vs-inferred reasoning, ontology discipline, ambiguity flagging, and self-QA.
02 Furniture
Polygon Segmentation Method: Polygon segmentation Tool: CVAT
Project objective
Trace the visible silhouettes of Chair, Table, and Sofa instances using polygon geometry. Curved contours were represented with sufficient vertices for accuracy while avoiding unnecessary point density. Hidden geometry was not reconstructed.
Results & QA
6 physical furniture instances segmented in 25:55. The scene included complex curves, foreground occlusion, and one chair split across five disconnected visible regions. Nine QA adjustments were made, with zero unresolved ambiguous cases.
Annotation insight
The exercise highlighted the distinction between a semantic object instance and the multiple annotation shapes sometimes required to represent its disconnected visible regions.
03 Precision Mask Segmentation
Method: Pixel/mask segmentation Tool: CVAT
Object: Round table with narrow legs and multiple foreground objects.
Original annotation time: approximately 35 minutes.
QA: 1 correction during high-zoom boundary review; 0 corrections during normal-view silhouette review.
Most difficult boundary: Thin table legs and narrow structural elements.
Key learning
Initial mask boundaries are not always pixel-accurate, particularly around narrow structures and complex contours.
High-zoom QA is necessary to identify and correct boundary errors before finalizing an annotation.