Abstract-- As global efforts to explore the Moon intensify, the need for high-quality 3D lunar maps becomes increasingly critical--particularly for long-distance missions such as NASA's Endurance mission concept, in which a rover aims to traverse 2,000 km across the South Pole-Aitken basin. Kaguya TC (T errain Camera) images, though globally available at 10 m/pixel, suffer from altitude inaccuracies caused by stereo matching errors and JPEG-based compression artifacts. This paper presents a method to improve the quality of 3D maps generated from Kaguya TC images, focusing on mitigating the effects of compression-induced noise in disparity maps. We analyze the compression behavior of Kaguya TC imagery, and identify systematic disparity noise patterns, especially in darker regions. In this paper, we propose an approach to enhance 3D map quality by reducing residual noise in disparity images derived from compressed images. Our experimental results show that the proposed approach effectively reduces elevation noise, enhancing the safety and reliability of terrain data for future lunar missions.

Weaknesses: I have no major concerns, but only remarks and suggestions for improvements. Although this is unambiguous in the experimental section, the abstract and introduction should clarify that the method is self-supervised from stereo pairs. There is a lot of confusion in the literature, because all monocular methods predict depth from a single image (by definition) but can be trained in different ways: from lidar supervision (full or partial), from stereo pairs (as is the case here), or from videos (a.k.a. Some of the authors' critique of related works (e.g., regarding dynamic objects) are only applicable to the SfM self-supervised scenario, as in the case of stereo-based self-supervised learning pairs of images are captured at the same time. Furthermore, the SfM case requires estimating the camera's ego-motion, which vastly complicates the self-supervised learning task (hence why the comparison is not entirely fair in my opinion).

We introduce a novel framework for training deep stereo networks effortlessly and without any ground-truth. By leveraging state-of-the-art neural rendering solutions, we generate stereo training data from image sequences collected with a single handheld camera. On top of them, a NeRF-supervised training procedure is carried out, from which we exploit rendered stereo triplets to compensate for occlusions and depth maps as proxy labels. This results in stereo networks capable of predicting sharp and detailed disparity maps. Experimental results show that models trained under this regime yield a 30-40% improvement over existing self-supervised methods on the challenging Middlebury dataset, filling the gap to supervised models and, most times, outperforming them at zero-shot generalization.

Apple has conquered the world of audio several times over. First with the iPod and iTunes, then with its ubiquitous AirPods - all of which changed the way the world listened to music and made phone calls. But the one device that failed to shift the dial was the HomePod, a voice-controlled Siri smart speaker that launched in 2018 and was discontinued in 2021 after lackluster sales. The smaller HomePod Mini remains on sale. But it never held a candle to the larger device, which sounded better than just about any rival when it launched in 2018, even if it lacked some of the'smart' features offered by rivals from Google and Amazon.

The great potential of unsupervised monocular depth estimation has been demonstrated by many works due to low annotation cost and impressive accuracy comparable to supervised methods. To further improve the performance, recent works mainly focus on designing more complex network structures and exploiting extra supervised information, e.g., semantic segmentation. These methods optimize the models by exploiting the reconstructed relationship between the target and reference images in varying degrees. However, previous methods prove that this image reconstruction optimization is prone to get trapped in local minima. In this paper, our core idea is to guide the optimization with prior knowledge from pretrained Flow-Net. And we show that the bottleneck of unsupervised monocular depth estimation can be broken with our simple but effective framework named FG-Depth. In particular, we propose (i) a flow distillation loss to replace the typical photometric loss that limits the capacity of the model and (ii) a prior flow based mask to remove invalid pixels that bring the noise in training loss. Extensive experiments demonstrate the effectiveness of each component, and our approach achieves state-of-the-art results on both KITTI and NYU-Depth-v2 datasets.

We study the effect of adversarial perturbations of images on deep stereo matching networks for the disparity estimation task. We present a method to craft a single set of perturbations that, when added to any stereo image pair in a dataset, can fool a stereo network to significantly alter the perceived scene geometry. Our perturbation images are "universal" in that they not only corrupt estimates of the network on the dataset they are optimized for, but also generalize to stereo networks with different architectures across different datasets. We evaluate our approach on multiple public benchmark datasets and show that our perturbations can increase D1-error (akin to fooling rate) of state-of-the-art stereo networks from 1% to as much as 87%. We investigate the effect of perturbations on the estimated scene geometry and identify object classes that are most vulnerable. Our analysis on the activations of registered points between left and right images led us to find that certain architectural components, i.e. deformable convolution and explicit matching, can increase robustness against adversaries. We demonstrate that by simply designing networks with such components, one can reduce the effect of adversaries by up to 60.5%, which rivals the robustness of networks fine-tuned with costly adversarial data augmentation.

Without a doubt, the Bang & Olufsen Beosound Level is the prettiest and best-sounding wireless streaming speaker I've encountered that operates on both AC and battery power. It holds true to the aesthetic spirit that's long driven this legendary Danish electronics brand, ever striving for a magical fusion of design and technology. But be forewarned, buying into this functional work of art isn't for the financially faint of heart: A single Beosound Level costs either $1,499 or $1,799, depending on which model you choose. You might need to rationalize this luxury indulgence as a long-term investment; like the pitch proffered for a Rolex timekeeper, or an exotic Euro sports car. And like those rare goods, the Beosound Level has a sensitive nature that in some ways demands a bit of coddling (more on that in a bit). At the heart of this device (and a few of B&O's other Wi-Fi-enabled speakers) is a new, modular circuit core called Mozart, which the manufacturer says can be swapped out for another if the onboard upgradability ever reaches its limit.

Summer is here and you're probably ready to party, see your friends, and mostly get back to yucking it up responsibly outdoors. But as plenty of people aren't comfortable venturing into the crowded world just yet, you're still going to spend time indoors. Either way, you'll want some tunes, and Sonos has made a speaker pretty well-suited for both environments. While the Sonos Roam has a few shortcomings, it's pretty impressive. But when traditional Bluetooth speakers have gotten pretty inexpensive, is it worth the hefty $169 price tag?

Smart speakers have become so ubiquitous lately that you most likely have more than one set up at home. Whether you're using smart speakers from Google, Amazon, or Apple, you can send audio to several speakers at once, configure them as stereo pairs, or even get your music to follow you from room to room. Google not long ago rebranded its Home speakers as Nest speakers, so you might have one or more of each--but they'll still work together no matter what the label says. Speaker management is handled through the Google Home app for Android or iOS. Open up the app and you'll see all your Nest speakers listed, together with any other connected smart home devices like Chromecasts and Nest cameras.

Amazon's Echo speakers represent the Next Big Thing in whole-home wireless audio: speakers that let you find and play music using only voice commands. And some of them sound quite good, although all of them are somewhat lacking when it comes to deep bass. To solve that problem, Amazon recently unveiled the Echo Sub, a subwoofer designed specifically to mate with the second-generation Echo and Echo Plus speakers, adding much deeper bass to their sonic palette. Is the Echo Sub worth adding $130 to your budget for smart speakers? The Echo Sub is quite diminutive as subwoofers go. The cylindrical, molded-plastic enclosure measures a mere 8 inches tall and 8.3 inches in diameter.