3D visual perception tasks, including 3D detection and map segmentation based on multi-camera images, are essential for autonomous driving systems. In this work, we present a new framework termed BEVFormer, which learns unified BEV representations with spatiotemporal transformers to support multiple autonomous driving perception tasks. In a nutshell, BEVFormer exploits both spatial and temporal information by interacting with spatial and temporal space through predefined grid-shaped BEV queries. To aggregate spatial information, we design spatial cross-attention that each BEV query extracts the spatial features from the regions of interest across camera views. For temporal information, we propose temporal self-attention to recurrently fuse the history BEV information. Our approach achieves the new state-of-the-art 56.9\% in terms of NDS metric on the nuScenes \texttt{test} set, which is 9.0 points higher than previous best arts and on par with the performance of LiDAR-based baselines. We further show that BEVFormer remarkably improves the accuracy of velocity estimation and recall of objects under low visibility conditions. The code is available at \url{https://github.com/zhiqi-li/BEVFormer}.
L04 Chinese 1070 Ampersand: Encountering China: A Performative Perspective on Chinese Culture and Identity This course examines the diversified and rich history of Chinese visual and performance cultures from the Chinese mainland, Taiwan, Hong Kong, and throughout the Chinese diaspora. A collaboration between the East Asian Languages and Cultures and Performing Arts departments, this course explores Chinese cultural narratives in relation to how they have been performed — on stage in traditional forms of dance-drama, on screen in film, and as lived in the practice of everyday life — from the late Imperial period to the present. It includes a practice component that introduces the students to movement disciplines such as Tai' Chi and opera, and it allows students to pursue creative assignments such as interview, stage plays, and filmmaking that demonstrate their developing knowledge of historical and contemporary Chinese culture. Building bridges of understanding between the United States and the Republic of China in Taiwan, the course will culminate in a spring break trip to Taiwan. This course is only for first-year, non-transfer students in the Ampersand: Encountering China program. Same as L61 FYP 107 Credit 3 units. A&S: AMP A&S IQ: LCD, SC BU: HUM, IS EN: H
We present an end-to-end deep learning architecture for depth map inference from multi-view images. In the network, we first extract deep visual image features, and then build the 3D cost volume upon the reference camera frustum via the differentiable homography warping. Next, we apply 3D convolutions to regularize and regress the initial depth map, which is then refined with the reference image to generate the final output. Our framework flexibly adapts arbitrary N-view inputs using a variance-based cost metric that maps multiple features into one cost feature. The proposed MVSNet is demonstrated on the large-scale indoor DTU dataset. With simple post-processing, our method not only significantly outperforms previous state-of-the-arts, but also is several times faster in runtime. We also evaluate MVSNet on the complex outdoor Tanks and Temples dataset, where our method ranks first before April 18, 2018 without any fine-tuning, showing the strong generalization ability of MVSNet.
When one captures images in low-light conditions, the images often suffer from low visibility. Besides degrading the visual aesthetics of images, this poor quality may also significantly degenerate the performance of many computer vision and multimedia algorithms that are primarily designed for high-quality inputs. In this paper, we propose a simple yet effective low-light image enhancement (LIME) method. More concretely, the illumination of each pixel is first estimated individually by finding the maximum value in R, G, and B channels. Furthermore, we refine the initial illumination map by imposing a structure prior on it, as the final illumination map. Having the well-constructed illumination map, the enhancement can be achieved accordingly. Experiments on a number of challenging low-light images are present to reveal the efficacy of our LIME and show its superiority over several state-of-the-arts in terms of enhancement quality and efficiency.
Understanding shifts in creative work will help guide AI’s impact on the media ecosystem The capabilities of a new class of tools, colloquially known as generative artificial intelligence (AI), is a topic of much debate. One prominent application thus far is the production of high-quality artistic media for visual arts, concept art, music, and literature, as well as video and animation. For example, diffusion models can synthesize high-quality images (1), and large language models (LLMs) can produce sensible-sounding and impressive prose and verse in a wide range of contexts (2). The generative capabilities of these tools are likely to fundamentally alter the creative processes by which creators formulate ideas and put them into production. As creativity is reimagined, so too may be many sectors of society. Understanding the impact of generative AI—and making policy decisions around it—requires new interdisciplinary scientific inquiry into culture, economics, law, algorithms, and the interaction of technology and creativity.
While large vision-language models (LVLMs) have demonstrated impressive capabilities in interpreting multi-modal contexts, they invariably suffer from object hallucinations (OH). We introduce HALC, a novel decoding algorithm designed to mitigate OH in LVLMs. HALC leverages distinct fine-grained optimal visual information in vision-language tasks and operates on both local and global contexts simultaneously. Specifically, HALC integrates a robust auto-focal grounding mechanism (locally) to correct hallucinated tokens on the fly, and a specialized beam search algorithm (globally) to significantly reduce OH while preserving text generation quality. Additionally, HALC can be integrated into any LVLMs as a plug-and-play module without extra training. Extensive experimental studies demonstrate the effectiveness of HALC in reducing OH, outperforming state-of-the-arts across four benchmarks.
In parallel with the dissemination of information technology, we note the persistence of frontiers within creative practices, in particular between the digital arts and the performing arts. Crossings of these frontiers brought to light the need for a common appropriation of digital issues. As a result of this appropriation, the AvatarStaging platform and its software dimension AKN_Regie will be described in their use to direct avatars on a mixed theatre stage. Developed with the Blueprint visual language within Epic Games' Unreal Engine, AKN_Regie offers a user interface accessible to non-programming artists. This feature will be used to describe two perspectives of appropriation of the tool: the Plugin perspective for these users and the Blueprint perspective for programming artists who want to improve the tool. These two perspectives are then completed by a C++ perspective that aligns AKN_Regie with the language with which the engine itself is programmed. The circulations between these three perspectives are finally studied by drawing on work on the ecology of collective intelligence.
Superconducting junctions with a ferromagnet as the weak link, where triplet correlations can transport supercurrents over a substantial distance, have been of long-standing interest. In this work, we study the triplet transport in planar La$_{0.7}$Sr$_{0.3}$MnO$_3$ (LSMO) nanowire Josephson junctions with NbTi superconducting contacts. By meticulous ion etching with an artificial Pt hard mask, the NbTi/LSMO bilayer is structured to form an LSMO bridge without damaging its top layer. We observe superconducting (critical) currents of the order of 10$^{9}$ A/m$^2$ in a junction with a length of 1.3 $μ$m, and distinguishing superconducting quantum interference (SQI) patterns when sweeping a magnetic field perpendicular ($B_\perp$) to the plane of the wire or parallel ($B_\parallel$) to the plane and along the wire. The observed Gaussian-shaped SQI pattern is attributed to the diffusive transport of triplet pairs in the LSMO. Our work demonstrates that combinations of oxide magnets with conventional ($s$-wave) alloy superconductors can be a promising new route to realizing superconducting spintronics.
The remake is a central notion in East Asian culture and applies to most of its cultural heritage with only limited changes. The kimono’s sleeves may vary but the idea of the dress itself remains unchanged, and the ceramics may feature a new type of glaze, but the inherent form tends to remain the same. A quick look at Chûshingura [The Loyal 47 Ronin] on IMDB resulted in thirty matches, the earliest production dating from 1910, with remakes in different iterations running up to 2022.
The internationalization of specialized healthcare emphasizes multiculturalism, requiring adaptable hospital spaces. Sant Joan de Déu (SJD), a leading pediatric hospital managed by a Christian order, has created a multifaith room for prayer and meditation in the main lobby of the Pediatric Cancer Center Barcelona (PCCB). This manuscript presents an unpublished case study, showing the research conducted for the design of the multireligious room and the process of its construction. The methodology includes a bibliographic review, architectural analysis of three meditation spaces, and in-depth interviews with stakeholders. This project highlights SJD’s commitment to blending care and design, emphasizing the humanization of hospital spaces. The triad of religion, public space, and society makes more sense here than ever before.
Now in its third edition, the After the Damages Advanced Training Summer School project on risk management and the impacts on the existing built heritage resulting from natural or man-made disasters consolidates both the didactic results and the interest and participation of national and international part- ners in the Academy’s activities.
The areas of investigation and com- parison, also on an international scale, have concerned, among other things, the effects of climate change on the conservation of cultural heritage. This has been done starting from an interdisciplinary and integrated approach to the issues related to emergency management, safety, and reconstruction following catastrophic events of a different nature. [read more]
Asier Aranzubía, Santiago Aguilar, José L. Castro-de-Paz
Las prácticas pertenecientes al género fantástico que se ruedan en la escuela de cine Madrid entre 1947 y 1976 conforman un corpus de obras con la suficiente envergadura como para obligarnos a repensar la historia de dicho género cinematográfico en nuestro país. En este artículo se estudia la producción fantástica del Instituto de Investigaciones y Experiencias Cinematográficas-Escuela Oficial de Cinematografía (IIEC-EOC) en el doble contexto del cine español de género y el de la literatura popular de temática fantástica que se edita en castellano durante el periodo franquista. En la segunda parte del artículo se analizan en detalle algunas de las películas más representativas de la producción fantástica del centro de enseñanza artística (especializado en cine) más importante que ha existido en España. En concreto, se presta atención a las adaptaciones de I am Legend(Richard Matheson), The Playground(Ray Bradbury) y The Prize of Peril(Robert Sheckley) que realizan, por este orden, Mario Gómez Martín, Pedro Olea y Francisco Montolío.
The article aims to find out whether the appearance of the images of The Lord of Good Hope with its rare iconography in Lithuania was influ enced by the family history of the famous geologist Ignacy Domeyko (Lith. Ignotas Domeika, 1802–1889), while emphasising the contribution of his son, Rev. Hernán Domeyko (1859–1931), towards the dissemination of this cult. The history and iconography of the image, as well as the development of the veneration tradition in Santiago de Chile and the Vilnius Archdiocese is explored. The authorship of the painting of Vilnius Cathedral is estab lished, the erroneous date of its creation established in historiography is corrected, and the votives belonging to the painting are identified.
Educating in junior visual arts courses is difficult and demanding labor. While it is implicated in many research, the complexity of the lived experience of secondary visual arts education has seldom been the topic of prolonged and synthesized study. In this work, the potential of practice as a term to analyse and portray secondary visual arts instruction is studied. A variety of practice theories are initially evaluated to uncover similar themes and concepts. From this conceptual basis, four theoretical principles are created as a framework to address the classroom based complexity of secondary visual arts instruction. A methodological design is then developed from this conceptual base. In conclusion, the prospect of applying this practice-based approach to the study of secondary visual arts classrooms is addressed in connection to empirical research performed with teachers as co-researchers.
While he was busy making The Other Side of the Wind, Orson Welles was working on finding new financial supporters that would grant him both money and real productive and creative autonomy. Among the interlocutors identified by Welles during the 1970s, the research ignored the interest shown by The Directors Company, a New Hollywood production company, little known to the general public due to its very short life cycle, but which, at the at the same time, was able to create some of the most original films of the period, such as Paper Moon (1973) by Peter Bogdanovich or The Conversation (1974) by Francis Ford Coppola. Thanks to the rediscovery of the Welles archive inside Museo Nazionale del Cinema in Turin, it is now possible to consult a series of documents capable to show the relationship between the director of Citizen Kane (1941) and one of the most original companies of the New Hollywood. It is a completely new series of letters, written after an unknown screening of F for Fake, which demonstrate, through the highly original perspective of Orson Welles, how close was the possibility that New Hollywood could produce The Other Side of the Wind. Some unpublished letters from the Welles fund of the National Cinema Museum in Turin highlight the relationship between Welles and the Directors Company, in relation to the troubled production of The Other Side of the Wind.
Arts-informed and arts-based methods are becoming more visible in qualitative psychological research. This study demonstrates how the use of visual images, through two visual methods photo-elicitation (PE) and photovoice (PV), can act as an innovative research tool for researchers. This paper focuses on the impacts of the visual images on the research process. A systematic search strategy was used to search 10 health and social science databases, with 2478 relevant articles retrieved, 197 articles were identified for review, and 75 articles met inclusion criteria. Qualitative synthesis and thematic analysis were selected to provide a flexible framework for addressing the research questions. The findings demonstrate the value of visual images being their ability to materialize bodies, social practices, and interactions between individuals and structures. These findings show that these two visual methods present researchers with an innovative process that can generate new insights and perspectives that are useful for studying aspects of gender relations (e.g., femininities and masculinities) and other aspects of social experience.