LMPT: Prompt Tuning with Class-Specific Embedding Loss for Long-tailed Multi-Label Visual Recognition

TL;DR

LMPT addresses long-tailed multi-label visual recognition by integrating prompt tuning with class-specific embedding loss and text-guided semantics. By freezing encoders and learning class-aware prompts, and by enforcing margins and weights that favor tail classes, the method aligns image features with caption-derived semantics across head and tail categories. The approach yields state-of-the-art results on VOC-LT and COCO-LT, outperforming LTML and zero-shot CLIP, and demonstrates the effectiveness of text supervision and distribution-aware learning in LTML tasks. This framework offers a practical, scalable way to leverage vision-language models for imbalanced multi-label settings with strong cross-modal semantic coupling.

Abstract

Long-tailed multi-label visual recognition (LTML) task is a highly challenging task due to the label co-occurrence and imbalanced data distribution. In this work, we propose a unified framework for LTML, namely prompt tuning with class-specific embedding loss (LMPT), capturing the semantic feature interactions between categories by combining text and image modality data and improving the performance synchronously on both head and tail classes. Specifically, LMPT introduces the embedding loss function with class-aware soft margin and re-weighting to learn class-specific contexts with the benefit of textual descriptions (captions), which could help establish semantic relationships between classes, especially between the head and tail classes. Furthermore, taking into account the class imbalance, the distribution-balanced loss is adopted as the classification loss function to further improve the performance on the tail classes without compromising head classes. Extensive experiments are conducted on VOC-LT and COCO-LT datasets, which demonstrates that our method significantly surpasses the previous state-of-the-art methods and zero-shot CLIP in LTML. Our codes are fully public at https://github.com/richard-peng-xia/LMPT.

Figures (5)

• Figure 1: The class distribution is long-tailed and the VLM compares image embeddings$\star$ to text embeddings$\bullet$$\blacksquare$$\blacktriangle$ of the class, which means the closer the distance between the embeddings of different modalities, the higher the probability that the category of the text embeddings matches the image. (a) Person and horse in the image belong to the head classes and the tail classes respectively. (b) Zero-Shot CLIP. (c) Exsiting Prompt Tuning w/o CSE loss. (d) LMPT (Ours) w/ CSE loss.
• Figure 2: Overview of the architecture of our proposed method. The color blocks are defined as shown in Fig. 1.
• Figure 3: The class margins (dotted lines) are enforced for generated samples by updating the decision boundary with respect to class margins.
• Figure 4: mAP performance of different methods w/o text supervision on two datasets. (a) VOC-LT. (b) COCO-LT.
• Figure 5: Example decisions from our model, CLIP, and ResNet.