KDEF-LM: A Facial Landmark Dataset with Emotion Labels Derived from the KDEF Dataset

Facial landmark features play a crucial role in emotion recognition, facial alignment, and human-computer interaction tasks. However, publicly available datasets that combine landmark coordinates with emotion labels are scarce. To address this, we introduce KDEF-LM, a curated and preprocessed facial landmark dataset derived from the Karolinska Directed Emotional Faces (KDEF) dataset. KDEF-LM contains 901 facial images annotated with 68 (x, y) landmark coordinates and corresponding emotion labels. The original 252\u00d7252 grayscale images have been resized to 48\u00d748 for computational efficiency. Images with undetectable faces were removed to ensure data quality. By bridging the gap between geometry-based facial features and emotion classification, KDEF-LM provides a valuable resource for researchers in affective computing and facial behavior analysis.