Recognizing continuous changes offers valuable insights into past historical events, supports current trend analysis, and facilitates future planning. This knowledge is crucial for a variety of fields, such as meteorology and agriculture, environmental science, urban planning and construction, tourism, and cultural preservation. Currently available datasets in the field of scene change understanding primarily concentrate on two main tasks: the detection of changed regions within a scene and the linguistic description of the change content. Existing datasets focus on recognizing discrete changes, such as adding or deleting an object from two images, and largely rely on artificially generated images. Consequently, the existing change understanding methods primarily focus on identifying distinct object differences, overlooking the importance of continuous, gradual changes occurring over extended time intervals. To address the above issues, we propose a novel benchmark dataset, STVchrono, targeting the localization and description of long-term continuous changes in real-world scenes. The dataset consists of 71,900 photographs from Google Street View API taken over an 18-year span across 50 cities all over the world. Our STVchrono dataset is designed to support real-world continuous change recognition and description in both image pairs and extended image sequences, while also enabling the segmentation of changed regions. We conduct experiments to evaluate state-of-the- art methods on continuous change description and segmentation, as well as multimodal Large Language Models for describing changes. Our findings reveal that even the most advanced methods lag human performance, emphasizing the need to adapt them to continuously changing real-world scenarios. We hope that our benchmark dataset will further facilitate the research of temporal change recognition in a dynamic world. The STVchrono dataset is available at STVchrono Dataset.

In outdoor environments, Vision-and-Language Navigation (VLN) requires an agent to rely on multi-modal cues from real-world urban environments and natural language instructions. While existing outdoor VLN models predict actions using a combination of panorama and instruction features, this approach ignores objects in the environment and learns data bias to fail navigation. According to our preliminary findings, most instances of navigation failure in previous models were due to turning or stopping at the wrong place. In contrast, humans intuitively frequently use identifiable objects or store names as reference landmarks, ensuring accurate turns and stops, especially in unfamiliar places. To address this insight gap, we propose an Object-Attention VLN (OAVLN) model that helps the agent focus on relevant objects during training and understand the environment better. Our model outperforms previous methods in all evaluation metrics under both seen and unseen scenarios on two existing benchmark datasets, Touchdown and map2seq.

Visual inspection robots used in factories and outdoor environments require the ability to accurately recognize visual differences between similar objects and further verbalize the recognition results to present the differences to humans. Despite the application of Large Language Models (LLMs) and multimodal LLMs across various domains, our research highlights their insufficiency in verbalizing nuanced differences across images. To address this, we leveraged LLMs and image generation AI to develop a dataset aimed at assessing difference recognition capabilities. We introduced two novel tasks using this dataset: selecting images based on their visual differences and a conditional difference captioning task, and evaluated existing Vision-Language Models (VLMs) on these tasks. Our findings reveal that advanced models like GPT-4V can describe subtle differences with comparative expressions, yet they fall short of matching human performance across all attributes. This discrepancy between model and human recognition, especially in identifying easily discernible differences, suggests that most current models lack the ability to directly compare image pairs for difference detection. Consequently, we propose a new model that incorporates an image-text similarity approach in the difference recognition task, showing superior performance over existing models, including GPT-4V. Our dataset and findings will contribute to advancements in differencing objects and improve robotic applications in visual inspection and object picking. The dataset is available at DICTA challenge page.

Abstract Hazy images suffer from two problems. The low contrast can be enhanced by estimating a transmission layer, and the colour cast can be restored by estimating an airlight. These two variables, together with the albedo layer, are the constitutive elements of a hazy image. The resulting quality of dehazed images is inextricably linked to the accurate estimation of these components. However, it is ill‐posed to decompose these variables from a single image. As such, this paper presents an innovative algorithm intended to facilitate the optimal decomposition of a hazy image. Using the Markov random field model, an optimal framework is established that allows the simultaneous estimation of the three components across the three‐colour channels. To improve the visual quality, three improvements are proposed in the variational solution for the optimal components. The dehazed result is recomposed from the components with the transmission enhanced to circumvent any potential artefacts or information loss. Extensive experiments on natural images corroborate that the proposed algorithm outperforms state‐of‐the‐art dehazing methods, both qualitatively and quantitatively.

With the improvement of digital image display technology, the “secondary imaging” caused by digital cameras is also gradually popularized, and the quality of the recaptured image formed by this imaging is also getting higher and higher, and this kind of high-quality fake image has caused great threat to digital images security. We propose a neural network-based recaptured image identification algorithm and use the difference between two types of images to build the identification algorithm in the frequency domain. The algorithm uses filtering to obtain the feature images which are the high-frequency and low-frequency filtering images, in order to further distinguish the image differences, the direction of the filtered image obtained from high-frequency images, each direction of the filtered image contains high-frequency information at different angles, and the low-frequency image is downsampled. At the same time, the low-frequency image is downsampled to obtain a multi-scale filtered image. The algorithm extracts the features from previous images as the feature values for classification, and finally uses neural networks for classification to obtain the classification results, and these prove that the algorithm presented is able to differentiate the recaptured images effectively in this paper.

This paper proposes a new visual reasoning formulation that aims at discovering changes between image pairs and their temporal orders. Recognizing scene dynamics and their chronological orders is a fundamental aspect of human cognition. The aforementioned abilities make it possible to follow step-by-step instructions, reason about and analyze events, recognize abnormal dynamics, and restore scenes to their previous states. However, it remains unclear how well current AI systems perform in these capabilities. Although a series of studies have focused on identifying and describing changes from image pairs, they mainly consider those changes that occur synchronously, thus neglecting potential orders within those changes. To address the above issue, we first propose a visual transformation graph structure for conveying order-aware changes. Then, we benchmarked previous methods on our newly generated dataset and identified the issues of existing methods for change order recognition. Finally, we show a significant improvement in order-aware change recognition by introducing a new model that explicitly associates different changes and then identifies changes and their orders in a graph representation.

With the rapid development of digital phones, the digital image forensics system in current times has had a great impact. It will lead to a serious threat for us, and especially the emergence of the recaptured image makes the existing digital image forensics algorithm invalid. So, it needs an effective image detection algorithm for us to identify recaptured images. In this paper, a new detection algorithm of the recaptured image is presented based on gray level co-occurrence matrix by analyzing the differences between the real and recaptured images. In order to analyze the differences, a new image evaluation model was put forward in this paper, which is called image variance ratio. Firstly, the algorithm proposed extracted high-frequency and low-frequency information of images by wavelet transform, based on which we calculated the relative gray level co-occurrence matrices. Secondly, the features of gray level co-occurrence matrix were extracted. At last, the recaptured image was classified by the support vector machine according to the features. The experimental results showed the algorithm proposed can not only effectively identify the recaptured image obtained from different media but also have better identification rate.

With the rapid development of digital cameras and smart phones, the image identification system in current times will be of a great impact. This will cause the form of image information to increase serious security issues. Especially, the emergence of the recaptured image makes conventional digital image forensics algorithm invalid. Therefore, a new image forensics algorithm is urgently needed to identify the recaptured image. In this paper, a new recaptured image identifying algorithm is put forward based on wavelet transformation and noise analysis by analyzing the differences between the real and recaptured images generated in the imaging process. First, the proposed algorithm extracts mean value, variance and skewness as wavelet characteristic from the high-frequency images and low-frequency images by wavelet transformation. Meanwhile, the proposed algorithm analyzes the noise image by means of local binary pattern to extract noise characteristic. Finally, the support vector machine is applied to classify the recaptured image with wavelet characteristics and noise characteristics. The results show the presented method can not only identify the recaptured image obtained from different media but also have better identification rate, and the dimension of the characteristic vector is also lower than those obtained by other algorithms.

It is now well expected that low-power sensor networks will soon be deployed for a wide variety of applications. These networks could potentially have millions of nodes spread in complex indoor/outdoor environments. One of the major deployment challenges under such diverse communication environments is providing reliable communication links to those low cost and/or battery-powered sensor nodes. Over the past few years, research in physical (PHY)-layer has demonstrated promising progresses on link reliability and energy efficiency. In modern medium access control (MAC) design, energy efficiency has become one of the key requirements and is still a hot research topic. In this overview, we provide a broad view encompassing both PHY- and MAC-layer techniques in the field of sensor networks with a focus on link reliability and energy efficiency. We review work in systems employing various PHY techniques in spatial diversity, energy efficient modulation, packet recovery, and data fusion, as well as MAC protocols in contention-based duty cycling, contention-free duty cycling, and hybrid duty cycling. The latest developments in cross-layer MAC designs that leverage PHY-layer techniques are presented. We also provide a synopsis of recent development and evolution of sensor network applications in industrial communications.

The use of asynchronous duty cycling at the MAC layer affords substantial energy savings in wireless networks. This technique is widely used in sensor networks and other types of wireless networks such as ad hoc networks. With asynchronous duty cycling, each node switches alternately between sleeping and active states; each node waking up asynchronously reduces network contention and wireless collisions caused by nodes waking up simultaneously, but also can have undesirable effects on higher layer protocols. In this paper, we study the problem of on-demand route discovery in asynchronous duty-cycling wireless networks and present four optimizations for such route discovery: Delayed Selection, Duty-Cycled Selection, Reply Updating, and Adaptive Backoff. Through detailed ns-2 simulations, we show that, without these optimizations, the routes discovered in asynchronous duty-cycling networks can be over 50% longer than the theoretical shortest routes and can have an ETX 90% larger than the ETX of the optimal routes. With only simple changes made at the MAC or network layers, our optimizations enabled nodes to substantially improve discovered routes, finding routes that were only 0.2% longer than the theoretical shortest routes or routes with an ETX only 9% larger than the ETX of the theoretical optimal-ETX routes, while also reducing route discovery latency and node energy consumption.

This paper presents PW-MAC (Predictive-Wakeup MAC), a new energy-efficient MAC protocol based on asynchronous duty cycling. In PW-MAC, nodes each wake up to receive at randomized, asynchronous times. PW-MAC minimizes sensor node energy consumption by enabling senders to predict receiver wakeup times; to enable accurate predictions, PW-MAC introduces an on-demand prediction error correction mechanism that effectively addresses timing challenges such as unpredictable hardware and operating system delays and clock drift. PW-MAC also introduces an efficient prediction-based retransmission mechanism to achieve high energy efficiency even when wireless collisions occur and packets must be retransmitted. We evaluate PW-MAC on a testbed of MICAz motes and compare it to X-MAC, WiseMAC, and RI-MAC, three previous energy-efficient MAC protocols, under multiple concurrent multihop traffic flows and under hidden-terminal scenarios and scenarios in which nodes have wakeup schedule conflicts. In all experiments, PW-MAC significantly outperformed these other protocols. For example, evaluated on scenarios with 15 concurrent transceivers in the network, the average sender duty cycle for X-MAC, WiseMAC, and RI-MAC were all over 66%, while PW-MAC's average sender duty cycle was only 11%; the delivery latency for PW-MAC in these scenarios was less than 5% that for WiseMAC and X-MAC. In all experiments, PW-MAC maintained a delivery ratio of 100%.

Medium access control (MAC) protocols for wireless sensor networks face many challenges, including energy-efficient operation and robust support for varying traffic loads, in spite of effects such as wireless interference or even possible wireless jamming attacks. This paper presents the design and evaluation of the EM-MAC (Efficient Multichannel MAC) protocol, which addresses these challenges through the introduction of novel mechanisms for adaptive receiver-initiated multichannel rendezvous and predictive wake-up scheduling. EM-MAC substantially enhances wireless channel utilization and transmission efficiency while resisting wireless interference and jamming by enabling every node to dynamically optimize the selection of wireless channels it utilizes based on the channel conditions it senses, without use of any reserved control channel. EM-MAC achieves high energy efficiency by enabling a sender to predict the receiver's wake-up channel and wake-up time Implemented in TinyOS on MICAz motes, EM-MAC substantially outperformed other MAC protocols studied. EM-MAC maintained the lowest sender and receiver duty cycles, the lowest packet delivery latency, and 100% packet delivery ratio across all experiments. Our evaluation includes single-hop and multihop flows, as well as experiments with heavy ZigBee interference, constant ZigBee jamming, and Wi-Fi interference.

We present Texas Instruments' implementation of IEEE 802.15.4e protocols (CSL and TSCH) in a multi-hop mesh. At the MAC layer, we discuss techniques which reduce radio duty-cycle in our CSL implementation. Further, we propose RPL adaptation, to operate the network in 2 phases, one with small DIO timer and another with large DIO timer to minimize the impact of RPL broadcasts. Our demo shows 10x reduction in radio power consumption resulting from careful implementation of our MAC and Routing layer techniques using TelosB motes running Contiki OS.

Sensor networks are increasingly available in home and industry automation applications. The use of battery-powered wireless sensor nodes eases network deployment, but the dynamic wireless medium, along with energy and bandwidth constraints, poses a great challenge to achieve a robust sensor network only with these nodes.

Traditional contention-based random-access wireless MAC protocols such as IEEE 802.11 DCF are designed for single-hop wireless networks and do not perform well in multi-hop scenarios due to inefficiency in their medium reservation mechanisms. In this paper, we introduce a new MAC protocol, called EMAC, which improves the efficiency of wireless medium reservation for general asynchronous multi-hop wireless networks. By exploiting limited routing information at the MAC layer, EMAC enables multiple asynchronous stations along a delivery path to cooperate in their random medium access. In particular, a control frame can travel across a multi-hop path composed of asynchronous nodes, and make synchronized medium reservations for an upcoming data frame transmission. This distributed cooperation at the MAC layer can greatly improve the medium reservation efficiency by reducing intra-flow contentions. Simulation results show that EMAC can improve end-to-end throughput compared to IEEE 802.11 DCF in chain scenarios by up to 85% and can avoid throughput starvation in cross scenarios.

The use of asynchronous duty-cycling in wireless sensor network MAC protocols is common, since it can greatly reduce energy consumption and requires no clock synchronization. However, existing systems using asynchronous duty-cycling do not efficiently support broadcast-based communication that may be used, for example, in route discovery or in network-wide queries or information dissemination. In this paper, we present the design and evaluation of ADB (Asynchronous Duty-cycle Broadcasting), a new protocol for efficient multihop broadcast in wireless sensor networks using asynchronous duty-cycling. ADB differs from traditional multihop broadcast protocols that operate above the MAC layer, in that it is integrated with the MAC layer to exploit information only available at this layer. Rather than treating the data transmission from a node to all of its neighbors as the basic unit of progress for the multihop broadcast, ADB dynamically optimizes the broadcast at the level of transmission to each individual neighbor of a node, as the neighbors asynchronously wakeup. We evaluate ADB both through ns-2 simulations and through measurements in a testbed of MICAz motes using TinyOS, and compare its performance to multihop broadcast based on X-MAC and on RI-MAC. In both evaluations, ADB substantially reduced energy consumption, network load, and delivery latency compared to other protocols, while achieving over 99% delivery ratio.

Duty cycling is a widely used mechanism in wireless sensor networks (WSNs) to reduce energy consumption due to idle listening, but this mechanism also introduces additional latency in packet delivery. Several schemes have been proposed to mitigate this latency, but they are mainly optimized for light traffic loads. A WSN, however, could often experience bursty and high traffic loads, such as due to broadcast or convergecast traffic. In this paper, we present a new MAC protocol, called Demand Wakeup MAC (DW-MAC), that introduces a new low-overhead scheduling algorithm that allows nodes to wake up on demand during the Sleep period of an operational cycle and ensures that data transmissions do not collide at their intended receivers. This demand wakeup adaptively increases effective channel capacity during an operational cycle as traffic load increases, allowing DW-MAC to achieve low delivery latency under a wide range of traffic loads including both unicast and broadcast traffic. We compare DW-MAC with S-MAC (with and without adaptive listening) and with RMAC using ns-2 and show that DW-MAC outperforms these protocols, with increasing benefits as traffic load increases. For example, under high unicast traffic load, DW-MAC reduces delivery latency by 70% compared to S-MAC and RMAC, and uses only 50% of the energy consumed with S-MAC with adaptive listening. Under broadcast traffic, DW-MAC reduces latency by more than 50% on average while maintaining higher energy efficiency.

The problem of idle listening is one of the most significant sources of energy consumption in wireless sensor nodes, and many techniques have been proposed based on duty cycling to reduce this cost. In this paper, we present a new asynchronous duty cycle MAC protocol, called Receiver-Initiated MAC (RI-MAC), that uses receiver-initiated data transmission in order to efficiently and effectively operate over a wide range of traffic loads. RI-MAC attempts to minimize the time a sender and its intended receiver occupy the wireless medium to find a rendezvous time for exchanging data, while still decoupling the sender and receiver's duty cycle schedules. We show the performance of RI-MAC through detailed ns-2 simulation and through measurements of an implementation in TinyOS in a testbed of MICAz motes. Compared to the prior asynchronous duty cycling approach of X-MAC, RI-MAC achieves higher throughput, packet delivery ratio, and power efficiency under a wide range of traffic loads. Especially when there are contending flows, such as bursty traffic or transmissions from hidden nodes, RI-MAC significantly improves throughput and packet delivery ratio. Even under light traffic load for which X-MAC is optimized, RI-MAC achieves the same high performance in terms of packet delivery ratio and latency while maintaining comparable power efficiency.

A major barrier for the adoption of wireless mesh networks is severe limits on throughput. In this paper, we apply superposition coding to substantially improve network capacity of large, dense wireless mesh networks. Superposition coding is a physical layer technique that allows a transmitter to simultaneously send independent packets to multiple receivers. While superposition coding has been studied extensively by the physical layer community, we present the first design of practical and effective MAC protocols to take advantage of superposition coding in wireless mesh networks. Extensive evaluations show that superposition coding can be a practical method to increase the throughput of large, dense wireless mesh networks. Specifically, in a mesh network with 2 to 64 active receivers and one gateway, we show that our system can increase throughput up to 154%, with average gain ranging from 10% to 21%. When there are multiple gateways forming a mesh network, our system gains up to 98%, with average gain ranging from 24% to 46%. These results clearly demonstrate the potential benefits of our system. We also present results from an implementation of superposition coding using GNU Radio.