Objects365: A Large-Scale, High-Quality Dataset for Object Detection

By Shuai Shao, Zeming Li, Tianyuan Zhang, Chao Peng, Gang Yu, Xiangyu Zhang, Jing Li. 2019

1,074 citations recorded by the source corpus.

In this paper, we introduce a new large-scale object detection dataset, Objects365, which has 365 object categories over 600K training images. More than 10 million, high-quality bounding boxes are manually labeled through a three-step, carefully designed annotation pipeline. It is the largest object detection dataset (with full annotation) so far and establishes a more challenging benchmark for the community. Objects365 can serve as a better feature learning dataset for localization-sensitive tasks like object detection and semantic segmentation. The Objects365 pre-trained models significantly outperform ImageNet pre-trained models with 5.6 points gain (42 vs 36.4) based on the standard setting of 90K iterations on COCO benchmark. Even compared with much long training time like 540K iterations, our Objects365 pretrained model with 90K iterations still have 2.7 points gain (42 vs 39.3). Meanwhile, the finetuning time can be greatly reduced (up to 10 times) when reaching the same accuracy. Better generalization ability of Object365 has also been verified on CityPersons, VOC segmentation, and ADE tasks. The dataset as well as the pretrained-models have been released at www.objects365.org.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

How Much Can CLIP Benefit Vision-and-Language Tasks? (2021, methodology): “VinVL (Li et al., 2020) is an extreme scaleup of the region-based paradigm, which is pretrained on multiple object detection datasets, including MS COCO (Lin et al., 2014), OpenImages (Kuznetsova et al., 2020), Object365 (Shao et al., 2019), and Visual Genome (Krishna et al., 2017).”
Florence: A New Foundation Model for Computer Vision (2021, methodology): “Fine-tuning We evaluate fine-tuning on three popular object detection datasets: COCO (Lin et al., 2015), Object365 (Shao et al., 2019), and Visual Genome (Krishna et al., 2016).”
Pix2seq: A Language Modeling Framework for Object Detection (2021, methodology): “For pretraining on Objects365 dataset, we use similar settings as above with a few differences.”
A Simple Framework for Open-Vocabulary Segmentation and Detection (2023, methodology): “For object detection, we use Objects365 [42] (660k images for v1 and 1700k images for v2).”
Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection (2023, methodology): “As the O365 dataset [44] has (nearly3) covered all categories in COCO, we evaluate an O365 pre-trined DINO on COCO as a zero-shot baseline.”
GLIGEN: Open-Set Grounded Text-to-Image Generation (2023, methodology): “Specifically, we follow GLIP [34] and train on Object365 [58] and GoldG [34], which combines two grounding datasets: Flickr [49] and VG [31].”
ELEVATER: A Benchmark and Toolkit for Evaluating Language-Augmented Visual Models (2022, methodology): “• DyHead [8] represents a checkpoint for the traditional language-free object detector, where model is pre-trained on Object365 [53] without leveraging the category name information.”
Grounded Language-Image Pre-training (2021, methodology): “, Objects365 [47]), in practice, we can split the categories into multiple prompts during training and inference.”
Learning Open-World Object Proposals without Learning to Classify (2021, methodology): “Therefore, in this work we study the generalization setting of training on COCO, then testing on other datasets: RoboNet [15], Objects365 [48], EpicKitchens [14], and LVIS [24].”
Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentation (2020, methodology): “When combined with self-training using unlabeled COCO and unlabeled Objects365 [49] for pseudo-labeling, we see a further improvement of 2.5 box AP and 2.2 mask AP, resulting in a model with a strong performance of 57.3 box AP and 49.1 mask AP on COCO test-dev without test-time augmen-tations and mo”
Rethinking Pre-training and Self-training (2020, background): “Note that OID is found to not be helpful on COCO by pre-training in [41].”
VinVL: Making Visual Representations Matter in Vision-Language Models (2021, methodology): “In Figure 8, we report the mAP 50 of checkpoints from these 4 experiments on 4 validation sets: COCO with stuff (top left), Objects365 (top right), OpenImages (bottom left) and Visual Genome (1594 object classes, bottom right).”
Open-vocabulary Object Detection via Vision and Language Knowledge Distillation (2021, methodology): “We evaluate the transferability of ViLD on PASCAL VOC (Everingham et al., 2010), COCO (Lin et al., 2014), and Objects365 (Shao et al., 2019).”
VIVO: Surpassing Human Performance in Novel Object Captioning with Visual Vocabulary Pre-Training (2020, background): “2020), Objects365 (Shao et al. 2019), etc.”
End-to-End Semi-Supervised Object Detection with Soft Teacher (2021, methodology): “* indicates that models are pre-trained with Object365 [24] dataset.”
Swin Transformer V2: Scaling Up Capacity and Resolution (2021, methodology): “For our largest model experiments, we employ an additional detection pre-training phase using Object 365 v2 dataset [59], in-between the image classiﬁcation pre-training phase and the COCO ﬁne-tuning phase.”
MiniVLM: A Smaller and Faster Vision-Language Model (2020, methodology): “First, we use the Objects365 dataset to pretrain TEE before fine-tuning it on VG, which improves the CIDEr score by 1.7, indicating that better visual fea- tures contribute to better performance on VL tasks.”