Long-term localization with map compression based on solar information

This paper addresses the problem of optimizing communicated information among heterogeneous, resourceaware robot teams to facilitate their navigation. In such operations, a mobile robot compresses its local map to assist another robot in reaching a target within an uncharted environment. The primary challenge lies in ensuring that the map compression step balances network load while transmitting only the most essential information for effective navigation. We propose a communication framework that sequentially selects the optimal map compression in a task-driven, communicationaware manner. It introduces a decoder capable of iterative map estimation, handling noise through Kalman filter techniques. The computational speed of our decoder allows for a larger compression template set compared to previous methods, and enables applications in more challenging environments. Specifically, our simulations demonstrate a remarkable 98% reduction in communicated information, compared to a framework that transmits the raw data, on a large Mars inclination map and an Earth map, all while maintaining similar planning costs. Furthermore, our method significantly reduces computational time compared to the state-of-the-art approach.

Obtaining a compact representation of a large-size pointset map built by mapper robots is a critical issue for recent SLAM applications. This “map compression” problem is explored from a novel perspective of dictionary-based map compression in the paper. The primary contribution of the paper is proposal of an incremental scheme for simultaneous mapping and map-compression applications. An incremental map compressor is presented by employing a modified RANSAC map-matching scheme as well as the compact projection visual search. Experiments show promising results in terms of compression speed, compactness of data and structure, as well as an application to the compression distance.

Compression of depth maps is important for “texture plus depth” format of multiview images, which enables synthesis of novel intermediate views via depth-image-based rendering (DIBR) at decoder. Previous depth map coding schemes exploit unique depth data characteristics to compactly and faithfully reproduce the original signal. In contrast, since depth map is only a means to the end of view synthesis and not itself viewed, in this paper we explicitly manipulate depth values, without causing severe synthesized view distortion, in order to maximize representation sparsity in the transform domain for compression gain - we call this process transform domain spar-sification (TDS). Specifically, for each pixel in the depth map, we first define a quadratic penalty function, with minimum at ground truth depth value, based on synthesized view's distortion sensitivity to the pixel's depth value during DIBR. We then define an objective for a depth signal in a block as a weighted sum of: i) signal's sparsity in the transform domain, and ii) per-pixel synthesized view distortion penalties for the chosen signal. Given that sparsity (l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> -norm) is non-convex and difficult to optimize, we replace the l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> -norm in the objective with a computationally inexpensive weighted l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -norm; the optimization is then an unconstrained quadratic program, solvable via a set of linear equations. For the weighted l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> -norm to promote sparsity, we solve the optimization iteratively, where at each iteration weights are readjusted to mimic sparsity-promoting l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">τ</sub> -norm, 0 ≤ τ ≤ 1. Using JPEG as an example transform codec, we show that our TDS approach gained up to 1.7dB in rate-distortion performance for the interpolated view over compression of unaltered depth maps.

Summary for only given. Conventional digital video playback systems provide only limited user interactivity, mostly in the form of VCR-like controls. In this model, the temporal ordering and the spatial viewpoints of the video streams being viewed are completely determined at authoring time. In contrast, we have defined a form of interactive video called active video (see http://www.ecsl.cs.sunysb.edu//spl sim/anindya/avs/avs.html, 2002), which supports hyper-linking among related video sequences and interpolation of video sequences with neighboring viewpoints, to offer end users the additional flexibility of choosing the sequencing and the viewing angle (even virtual ones) at playback time. However, active video has a substantially higher storage and transmission cost due to multiple time-synchronized video streams capturing the dynamic scene and the pixel-level correspondence maps encoding the spatial association among the frame-pairs of all adjacent streams. The maps are interpolated at run-time to generate virtual views. We describe and evaluate the following three compression techniques that alleviate the storage and network transmission costs of active video: spatial video compression; lossy map compression; lossless map compression.

High-Definition (HD) maps are needed for robust navigation of autonomous vehicles, limited by the on-board storage capacity. To solve this, we propose a novel framework, Environment-Aware Normal Distributions Transform (EA-NDT), that significantly improves compression of standard NDT map representation. The compressed representation of EA-NDT is based on semantic-aided clustering of point clouds resulting in more optimal cells compared to grid cells of standard NDT. To evaluate EA-NDT, we present an open-source implementation that extracts planar and cylindrical primitive features from a point cloud and further divides them into smaller cells to represent the data as an EA-NDT HD map. We collected an open suburban environment dataset and evaluated EA-NDT HD map representation against the standard NDT representation. Compared to the standard NDT, EA-NDT achieved consistently at least 1.5x higher map compression while maintaining the same descriptive capability. Moreover, we showed that EA-NDT is capable of producing maps with significantly higher descriptivity score when using the same number of cells than the standard NDT.

Level-crossing approach to the compression of body surface potential maps

In this paper, we consider a novel image coding paradigm, termed semantically scalable coding. In the new paradigm, coded bitstream serves for multiple different semantic analysis tasks, and different tasks require different semantic granularities of the image. Thus, the bitstream is designed to be scalable in the sense that progressive decoding of the bitstream provides coarse-to-fine semantic granularities. As a concrete example, we consider the task of coarse-grained and fine-grained image classification. We present a method to compress the multiple deep feature maps that are intermediate representations of an image passing a trained deep network. The deep-layer feature maps can serve for coarse-grained image classification while the shallow-layer feature maps can serve for fine-grained image classification. Experimental results demonstrate the feasibility of the proposed method, as well as the advantage of the semantically scalable coding paradigm.

A new method for the lossy compression of segmentation maps is presented. In the first stage of the proposed algorithm, region boundary segments are coded, providing a coarse polygonal approximation of regions, as well as a complete representation of their adjacency. The second stage consists of the progressive refinement of these segments, until either the desired degree of accuracy is reached, or the target bit-rate is achieved. Rate-distortion curves demonstrate good lossy performance in the range of 0.3 to 0.6 bits per contour point.

We consider applications using depth-based image-based rendering (IBR), where the synthesis of arbitrary views occur at a remote location, necessitating the compression and transmission of depth maps. Traditional image compression has been designed to provide maximum perceived visual quality, and a direct application is sub-optimal for depth-map compression, since depth-maps are not directly viewed. In other words, the sensitivity of the rendering error depends on the image content as well as on the depth map, we propose two improvements to take this into account. Firstly, we consider region-of-interest (ROI) coding, where we identify those regions of the image where accurate depth is most crucial. Secondly, we reshape the dynamic range of the depth map. Our experiments show a significant improvement in coding gain (1.1 dB) and rendering quality when we integrated these two improvements into a standard JPEG-2000 coder.

In order to deal with the scaling problem of volumetric map representations we propose spatially local methods for high-ratio compression of 3D maps, represented as truncated signed distance fields. We show that these compressed maps can be used as meaningful descriptors for selective decompression in scenarios relevant to robotic applications. As compression methods, we compare using PCA-derived low-dimensional bases to non-linear auto-encoder networks and novel mixed architectures that combine both. Selecting two application-oriented performance metrics, we evaluate the impact of different compression rates on reconstruction fidelity as well as to the task of map-aided ego-motion estimation. It is demonstrated that lossily compressed distance fields used as cost functions for ego-motion estimation, can outperform their uncompressed counterparts in challenging scenarios from standard RGB-D data-sets.

The efficient compression of depth maps is becoming more and more important. We present a novel codec specifically suited for this task. In the encoding step we segment the image and extract between-pixel contours. Subsequently we optimise the grey values at carefully selected mask points, including both hexagonal grid locations as well as freely chosen points. We use a chain code to store the contours. For the decoding we apply a segment-based homogeneous diffusion inpainting. The segmentation allows parallel processing of the individual segments. Experiments show that our compression algorithm outperforms comparable methods such as JPEG or JPEG2000, while being competitive with HEVC (High Efficiency Video Coding).

Convolutional Neural Networks (CNNs) are becoming a fundamental tool for machine learning. High performance and energy efficiency are of great importance for deployments of CNNs in many embedded applications. Energy consumption during CNN processing is dominated by memory access and since large networks do not fit in on-chip storage, they require expensive DRAM access. This paper introduces an universal Output Stream Manager (OSM) which can be used to compress and format data coming from a CNN accelerator and reduce external memory access. The OSM exploits the sparsity of data and implements two Zero-Run Length encoding algorithms and can be easily reconfigured to optimize usage for different CNN layers.

LatentMap: Effective auto-encoding of density maps for spatiotemporal data visualizations

Prospective 3D video transmission systems that would use compression of both multiview video and depth maps was dealt with in this paper. To get the optimal synthesized view quality with the transmission bit rate being constant, it firstly analyzed the overall behavior of combined coding and view synthesis, color and depth data with different coding qualities were combined for intermediate view synthesis. Then, individual coding results for color and depth data were evaluated in order to investigate, how well the synthesized view quality was at certain compression. Finally, a new joint coding method for multi-view video plus depth (MVD) was proposed by utilizing the optimal combinations of the quantization parameters of color videos and the corresponding depth maps. Simulation results demonstrated that the proposed method decreased the bit-rate by 12.72%~24.02% in comparison with JMVC and can be used to guide sequence-level texture/depth coding for 3D video compression. The proposed method has better performance in optimizing the tradeoff between bit-rate and visual quality.

We study the application of nonlinear max- and min-lifted wavelets to the compression of digital elevation maps (DEMs). Such wavelets have the desirable property that regional maximum or minimum values are preserved at coarse resolutions, making it possible with proper encoding to efficiently search for high/low points without decoding the entire file. In course of studying these nonlinear transforms, we first quantify the dynamic range expansion incurred by using them, and we then illustrate with an example how maximum values propagate to coarse scales. As a first step, we consider encoding the resulting coefficients using two popular algorithms: set partitioning in hierarchical trees (SPIHT) and stack-run. In doing this, we find that efficiently exploiting inter-scale redundancies in the transform is far more important for these wavelets than it is for the more commonly used (2,2)-integer wavelet (also called the 5/3 wavelet).