LAIP: Learning Local Alignment from Image-Phrase Modeling for Text-based Person Search

Text-based person search aims at retrieving images of a particular person based on a given textual description. A common solution for this task is to directly match the entire images and texts, i.e., global alignment, which fails to deal with discerning specific details that discriminate against appearance-similar people. As a result, some works shift their attention towards local alignment. One group matches fine-grained parts using forward attention weights of the transformer yet underutilizes information. Another implicitly conducts local alignment by reconstructing masked parts based on unmasked context yet with a biased masking strategy. All limit performance improvement. This paper proposes the Local Alignment from Image-Phrase modeling (LAIP) framework, with Bidirectional Attention-weighted local alignment (BidirAtt) and Mask Phrase Modeling (MPM) module.BidirAtt goes beyond the typical forward attention by considering the gradient of the transformer as backward attention, utilizing two-sided information for local alignment. MPM focuses on mask reconstruction within the noun phrase rather than the entire text, ensuring an unbiased masking strategy. Extensive experiments conducted on the CUHK-PEDES, ICFG-PEDES, and RSTPReid datasets demonstrate the superiority of the LAIP framework over existing methods.

Fully Sparse 3D Panoptic Occupancy Prediction

Occupancy prediction plays a pivotal role in the realm of autonomous driving. Previous methods typically constructs a dense 3D volume, neglecting the inherent sparsity of the scene, which results in a high computational cost. Furthermore, these methods are limited to semantic occupancy and fail to differentiate between distinct instances. To exploit the sparsity property and ensure instance-awareness, we introduce a novel fully sparse panoptic occupancy network, termed SparseOcc. SparseOcc initially reconstructs a sparse 3D representation from visual inputs. Subsequently, it employs sparse instance queries to predict each object instance from the sparse 3D representation. These instance queries interact with 2D features via mask-guided sparse sampling, thereby circumventing the need for costly dense features or global attention. Additionally, we have established the first-ever vision-centric panoptic occupancy benchmark. SparseOcc demonstrates its efficacy on the Occ3D-nus dataset by achieving a mean Intersection over Union (mIoU) of 26.0, while maintaining a real-time inference speed of 25.4 FPS. By incorporating temporal modeling from the preceding 8 frames, SparseOcc further improves its performance, achieving 30.9 mIoU without whistles and bells. Code will be made available.

SparseBEV: High-Performance Sparse 3D Object Detection from Multi-Camera Videos

Camera-based 3D object detection in BEV (Bird's Eye View) space has drawn great attention over the past few years. Dense detectors typically follow a two-stage pipeline by first constructing a dense BEV feature and then performing object detection in BEV space, which suffers from complex view transformations and high computation cost. On the other side, sparse detectors follow a query-based paradigm without explicit dense BEV feature construction, but achieve worse performance than the dense counterparts. In this paper, we find that the key to mitigate this performance gap is the adaptability of the detector in both BEV and image space. To achieve this goal, we propose SparseBEV, a fully sparse 3D object detector that outperforms the dense counterparts. SparseBEV contains three key designs, which are (1) scale-adaptive self attention to aggregate features with adaptive receptive field in BEV space, (2) adaptive spatio-temporal sampling to generate sampling locations under the guidance of queries, and (3) adaptive mixing to decode the sampled features with dynamic weights from the queries. On the test split of nuScenes, SparseBEV achieves the state-of-the-art performance of 67.5 NDS. On the val split, SparseBEV achieves 55.8 NDS while maintaining a real-time inference speed of 23.5 FPS. Code is available at https://github.com/MCG-NJU/SparseBEV.