Hough transform-based methods for detecting multiple objects use nonmaxima suppression or mode seeking to locate and distinguish peaks in Hough images. Such postprocessing requires the tuning of many parameters and is often fragile, especially when objects are located spatially close to each other. In this paper, we develop a new probabilistic framework for object detection which is related to the Hough transform. It shares the simplicity and wide applicability of the Hough transform but, at the same time, bypasses the problem of multiple peak identification in Hough images and permits detection of multiple objects without invoking nonmaximum suppression heuristics. Our experiments demonstrate that this method results in a significant improvement in detection accuracy both for the classical task of straight line detection and for a more modern category-level (pedestrian) detection problem.

No abstract

We present Adaptive Instance Selection network architecture for class-agnostic instance segmentation. Given an input image and a point (x, y), it generates a mask for the object located at (x, y). The network adapts to the input point with a help of AdaIN layers [13], thus producing different masks for different objects on the same image. Adap-tIS generates pixel-accurate object masks, therefore it accurately segments objects of complex shape or severely occluded ones. AdaptIS can be easily combined with standard semantic segmentation pipeline to perform panoptic segmentation. To illustrate the idea, we perform experiments on a challenging toy problem with difficult occlusions. Then we extensively evaluate the method on panoptic segmentation benchmarks. We obtain state-of-the-art results on Cityscapes and Mapillary even without pretraining on COCO, and show competitive results on a challenging COCO dataset. The source code of the method and the trained models are available at https://github.com/saicvul/adaptis.

To detect multiple objects of interest, the methods based on Hough transform use non-maxima supression or mode seeking in order to locate and to distinguish peaks in Hough images. Such postprocessing requires tuning of extra parameters and is often fragile, especially when objects of interest tend to be closely located. In the paper, we develop a new probabilistic framework that is in many ways related to Hough transform, sharing its simplicity and wide applicability. At the same time, the framework bypasses the problem of multiple peaks identification in Hough images, and permits detection of multiple objects without invoking nonmaximum suppression heuristics. As a result, the experiments demonstrate a significant improvement in detection accuracy both for the classical task of straight line detection and for a more modern category-level (pedestrian) detection problem.

We present a new optimization based parsing framework for the geometric analysis of a single image coming from a man-made environment. This framework models the scene as a composition of geometric primitives spanning different layers from low level (edges) through mid-level (lines segments, lines and vanishing points) to high level (the zenith and the horizon). The inference in such a model thus jointly and simultaneously estimates (a) the grouping of edges into the line segments, (b) the grouping of line segments into the straight lines, (c) the grouping of lines into parallel families, and (d) the positioning of the horizon and the zenith in the image. Such a unified treatment means that the uncertainty information propagates between the layers of the model. This is in contrast to most previous approaches to the same problem, which either ignore the middle levels (line segments or lines) all together, or use the bottom-up step-by-step pipeline.For the evaluation, we consider a publicly available York Urban dataset of "Manhattan" scenes, and also introduce a new, harder dataset of 103 urban outdoor images containing many non-Manhattan scenes. The comparative evaluation for the horizon estimation task demonstrate higher accuracy and robustness attained by our method when compared to the current state-of-the-art approaches.