S2S-Sim: A Benchmark Dataset for Ship Cooperative 3D Object Detection

Abstract

The rapid development of vehicle cooperative 3D object-detection technology has significantly improved the perception capabilities of autonomous driving systems. However, ship cooperative perception technology has received limited research attention compared to autonomous driving, primarily due to the lack of appropriate ship cooperative perception datasets. To address this gap, this paper proposes S2S-sim, a novel ship cooperative perception dataset. Ship navigation scenarios were constructed using Unity3D, and accurate ship models were incorporated while simulating sensor parameters of real LiDAR sensors to collect data. The dataset comprises three typical ship navigation scenarios, including ports, islands, and open waters, featuring common ship classes such as container ships, bulk carriers, and cruise ships. It consists of 7000 frames with 96,881 annotated ship bounding boxes. Leveraging this dataset, we assess the performance of mainstream vehicle cooperative perception models when transferred to ship cooperative perception scenes. Furthermore, considering the characteristics of ship navigation data, we propose a regional clustering fusion-based ship cooperative 3D object-detection method. Experimental results demonstrate that our approach achieves state-of-the-art performance in 3D ship object detection, indicating its suitability for ship cooperative perception.

Towards Super Compressed Neural Networks for Object Identification: Quantized Low-Rank Tensor Decomposition with Self-Attention

“…In this section, we introduce how to apply the vector quantization to the low-rank convolution kernels U (3) , G, and U (4) .…”

“…The overall framework of QLTD Input: weights W ∈ R d×d×C in ×C out , the channel compression ratio of the Tucker decomposition r, the length of subvectors in vector quantization clustering d, the number of cluster centroids k, and the self-attention percentage s. 4) and B G , the initial weight of the self-attention module S U (3) , S U (4) and S G . 1: Obtain U (3) , G and U (4) by applying Tucker-2 decomposition to W ∈ R d×d×C in ×C out as Equation ( 4). 2: Split U (3) to sub-vectors by Ĉin = C in /d as Equation (6).…”

“…3: Similarly, obtain the codebook C U (4) b and code B U (4) of U (4) . 4: Similarly, obtain the codebook C G b and code B G of G. 5: Apply the same permutation to the output dimension for parent layers and the input dimension for children layers.…”

“…Deep convolutional neural networks (DCNNs) have realized great success in many computer vision tasks, such as object identification [1][2][3], object detection [4][5][6] and image segmentation [7][8][9]. The design of DCNNs is becoming increasingly intricate, which is accompanied by a simultaneous enhancement in their performance.…”

“…The self-attention modules are sparse and trainable. Additionally, a permutation and vector quantization approach is applied to U (3) , G and U (4) to further reduce parameter storage, which is explained in Sections 3.3 and 3.4 in detail.…”