SOLO: Segmenting Objects by Locations

By Xinlong Wang, Tao Kong, Chunhua Shen, Yuning Jiang, Lei Li. 2019

839 citations recorded by the source corpus.

We present a new, embarrassingly simple approach to instance segmentation in images. Compared to many other dense prediction tasks, e.g., semantic segmentation, it is the arbitrary number of instances that have made instance segmentation much more challenging. In order to predict a mask for each instance, mainstream approaches either follow the 'detect-thensegment' strategy as used by Mask R-CNN, or predict category masks first then use clustering techniques to group pixels into individual instances. We view the task of instance segmentation from a completely new perspective by introducing the notion of "instance categories", which assigns categories to each pixel within an instance according to the instance's location and size, thus nicely converting instance mask segmentation into a classification-solvable problem. Now instance segmentation is decomposed into two classification tasks. We demonstrate a much simpler and flexible instance segmentation framework with strong performance, achieving on par accuracy with Mask R-CNN and outperforming recent singleshot instance segmenters in accuracy. We hope that this very simple and strong framework can serve as a baseline for many instance-level recognition tasks besides instance segmentation.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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

Citations in the corpus

Global Pooling, More than Meets the Eye: Position Information is Encoded Channel-Wise in CNNs (2021, neutral): “…Islam et al. [16] performed a large-scale case study on border heuristics, including padding and canvas colors, and showed that positional information can hurt (e.g., texture recognition [5]) or help (e.g., semantic segmentation and instance segmentation [32]) performance depending on the task.”
K-Net: Towards Unified Image Segmentation (2021, methodology): “Therefore, we apply Dice loss [37] to handle this issue following previous works [45, 47, 48].”
Conditional Convolutions for Instance Segmentation (2020, methodology): “Wang et al. developed a simple FCN based instance segmentation method, showing competitive performance [22].”
End-to-End Video Instance Segmentation with Transformers (2020, background): “While signiﬁcant progress has been witnessed in instance segmentation of images [5,9,22,25–27], much less effort was spent on segmenting instances in videos.”
RTMDet: An Empirical Study of Designing Real-Time Object Detectors (2022, background): “Recent attempts perform instance segmentation in one stage with [4, 69] or without bounding boxes [76, 77, 96].”
Vision-based robotic grasping from object localization, object pose estimation to grasp estimation for parallel grippers: a review (2020, background): “SOLO [Wang et al., 2019f] introduces the notion of instance categories, which assigns categories to each pixel within an instance according to the instance’s location and size, and converts instance mask segmentation into a classificationsolvable problem.”
BoxInst: High-Performance Instance Segmentation with Box Annotations (2020, background): “The performance of instance segmentation has been significantly advanced by a number of recent successful methods [5, 6, 11, 15, 28, 30, 31].”
SOLQ: Segmenting Objects by Learning Queries (2021, background): “Regardless of the bounding boxes, SOLO [15] introduces the notion of "instance categories" and segments objects by locations.”
Interactive Object Segmentation With Inside-Outside Guidance (2022, neutral): “…a revolutionary advancement in semantic [9], [10], [11], [12], [13], [14], [15], [16], [17], [18] and instance segmentation [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29] for different domains, such as general scenes [3], [30], [31], autonomous driving [5], [32], [33],…”
Crossover Learning for Fast Online Video Instance Segmentation (2021, background): “Many prior works [25, 10, 14, 28, 9, 12, 5, 62, 56, 32, 6, 34] contribute a lot to the rapid developments in this ﬁeld.”
A survey on instance segmentation: state of the art (2020, neutral): “3% 2020 SOLO [136] 40.”
Spatial Feature Calibration and Temporal Fusion for Effective One-stage Video Instance Segmentation (2021, background): “One-stage instance segmentation approaches [42, 34, 7, 45, 38, 5, 40, 21] recently are proposed to keep the trade-off between speed and performance.”
FaPN: Feature-aligned Pyramid Network for Dense Image Prediction (2021, neutral): “There is a plethora of encoder-decoder methods [12, 13, 18, 23, 37, 43, 45] proposed for different dense image prediction tasks.”
PENet: Towards Precise and Efficient Image Guided Depth Completion (2021, methodology): “However, encoding pixel coordinates (B4+CCL) may slightly hurt the performance, contradicting to its performance in other position sensitive vision tasks [23], [24].”
Instances as Queries (2021, methodology): “In addition to R-CNN based methods, YOLACT [3, 2], SipMask [6], CondInst [46] and SOLO [51, 52] build one-stage instance segmentation framework on the top of one-stage framework, achieving comparable results with favorable inference speed.”
FoveaBox: Beyound Anchor-Based Object Detection (2019, background): “Also, there are several works trying to extend the anchor-free concept into instance segmentation [59], [60].”
Images Speak in Images: A Generalist Painter for In-Context Visual Learning (2022, methodology): “To address this optimization issue, we follow SOLO [42] to assign the color of each instance mask according to the absolute position of its center in the image.”