Explore the vast range of titles available.

In the last decade, deep learning has enjoyed its spotlight as the game-changing addition to smart farming and precision agriculture. Such development has been predominantly observed in developed countries, while on the other hand, in developing countries most farmers especially ones with smallholder farms have not enjoyed such wide and deep adoption of this new technologies. In this paper we attempt to improve the image classification part of smart farming and precision agriculture. Agricultural commodities tend to possess certain textural details on their surfaces which we attempt to exploit. In this work, we propose a deep learning based approach called Selective Context Adaptation Network (SCANet). SCANet performs feature enhancement strategy by leveraging level-wise information and employing context selection mechanism. In exploiting contextual correlation feature of the crop images our proposed approach demonstrates the effectiveness of the context selection mechanism. Our proposed scheme achieves 88.72% accuracy and outperforms the existing approaches. Our model is evaluated on the cocoa bean dataset constructed from the real cocoa bean industry scene in Indonesia.

Plant lesion segmentation aims to delineate disease regions at the pixel level to support early diagnosis, severity assessment, and targeted intervention in precision agriculture. However, the task remains challenging due to large variations in lesion scale-ranging from minute incipient spots to coalesced regions-and ambiguous, low-contrast boundaries that blend into healthy tissue. We present GARDEN, a Gradient-guided boundary-Aware Region-Driven Edge-refiNement network that unifies multi-scale context modeling with selective long-range boundary refinement. Our approach integrates a Multi-Scale Context Aggregation (MSCA) module to harvest contextual cues across diverse receptive fields, forming scale-consistent lesion priors to improve sensitivity to tiny lesions. Additionally, we introduce a Boundary-aware Selective Scanning (BASS) module conditioned on a Gradient-Guided Boundary Predictor (GGBP). This module produces an explicit boundary prior to steer a Mamba-based 2D selective scan, allocating long-range reasoning to boundary-uncertain pixels while relying on local evidence in confident interiors. Validated across two public plant disease datasets, GARDEN achieves state-of-the-art results on both overlap and boundary metrics. Specifically, the model demonstrates pronounced gains on small lesions and boundary-ambiguous cases. Qualitative results further show sharper contours and reduced spurious responses to illumination and viewpoint changes compared to existing methods. By coupling scale robustness with boundary precision in a single architecture, GARDEN delivers accurate and reliable plant lesion segmentation. This method effectively addresses key challenges in the field, offering a robust solution for automated disease analysis under challenging real-world conditions.

In rich visual environments, humans have to adjust their attentional control settings in various ways, depending on the task. Especially if the environment changes dynamically, it remains unclear how observers adapt to these changes. In two experiments (online and lab-based versions of the same task), we investigated how observers adapt their target choices while searching for color singletons among shape distractor contexts that changed over trials. The two equally colored targets had shapes that differed from each other and matched a varying number of distractors. Participants were free to select either target. The results show that participants adjusted target choices to the shape ratio of distractors: even though the task could be finished by focusing on color only, participants showed a tendency to choose targets matching with fewer distractors in shape. The time course of this adaptation showed that the regularities in the changing environment were taken into account. A Bayesian modeling approach was used to provide a fine-grained picture of how observers adapted their behavior to the changing shape ratio with three parameters: the strength of adaptation, its delay relative to the objective distractor shape ratio, and a general bias toward specific shapes. Overall, our findings highlight that systematic changes in shape, even when it is not a target-defining feature, influence how searchers adjust their attentional control settings. Furthermore, our comparison between lab-based and online assessments with this paradigm suggests that shape is a good choice as a feature dimension in adaptive choice online experiments.

Evidence across a wide variety of attention paradigms shows that environmental cues can trigger adjustments to ongoing priorities for attending to relevant and irrelevant information. This context-specific control over attention suggests that cognitive control can be both automatic and flexible. For instance, in selective attention tasks, congruency effects are larger for items that appear in a context associated with infrequent conflict than in a context associated with frequent conflict. Because the to-be-presented context cannot be predicted or prepared for in advance, attention is assumed to be rapidly updated on-the-fly, triggered by the currently presented context. Context-specific control exemplifies how learning and memory processes can influence attention to enable cognitive flexibility. However, what determines the use of previously learned associations remains unclear. In the current study, we examined whether task relevance would influence the learning and use of context cues in a flanker task. Using a secondary counting task, context dimensions associated with differing levels of conflict were made task-relevant or -irrelevant across the experiment. In short, we found that making new contextual information task-relevant caused participants to suppress a previously learned context-attention association and adopt a new context-specific control strategy--all without changing the experimental stimuli. Furthermore, we found participants did not spontaneously learn about context-specific proportion manipulations (Experiment 2) and explicit instructions were insufficient for producing context-specific effects (Experiment 3). These results suggest that task relevance is a key determinant of context-specific control. All data, analyses, article preparation, and experimental design code is available at https://osf.io/ztcyb/. Les données recueillies dans le cadre d'une grande variété de paradigmes d'attention montrent que les indices environnementaux peuvent déclencher des ajustements des priorités en cours pour traiter des informations pertinentes et non pertinentes. Ce contrôle de l'attention spécifique au contexte suggère que le contrôle cognitif peut être à la fois automatique et souple. Par exemple, dans les tâches d'attention sélective, les effets de congruence sont plus importants pour les items qui apparaissent dans un contexte associé à des conflits peu fréquents que dans un contexte associé à des conflits fréquents. Étant donné que le contexte à présenter ne peut être prédit ou préparé à l'avance, on suppose que l'attention sera rapidement mise à jour sur-le-champ, déclenchée par le contexte actuel. Le contrôle spécifique au contexte illustre comment les processus d'apprentissage et de mémoire peuvent influencer l'attention pour permettre une flexibilité cognitive. Toutefois, ce qui détermine l'utilisation des associations acquises auparavant reste flou. Dans l'étude en cours, nous avons examiné si la pertinence de la tâche avait une incidence sur l'apprentissage et l'utilisation d'indices de contexte dans une tâche d'accompagnement. En utilisant une tâche de comptage secondaire, les dimensions contextuelles associées aux différents niveaux de conflit ont été rendues pertinentes ou non pertinentes à la tâche tout au long de l'expérience. En bref, nous avons découvert que le fait de rendre la nouvelle information contextuelle pertinente à la tâche a incité les participants à supprimer une association contexte-attention apprise précédemment et à adopter une nouvelle stratégie de contrôle spécifique au contexte - tout cela sans changer les stimuli expérimentaux. De plus, nous avons constaté que les participants n'avaient pas spontanément appris au sujet des manipulations de proportions spécifiques au contexte (expérience 2) et que les instructions explicites étaient insuffisantes pour produire des effets spécifiques au contexte (expérience 3). Ces résultats suggèrent que la pertinence de la tâche est un déterminant clé du contrôle propre au contexte. Toutes les données, les analyses, la préparation des articles et le code de conception expérimentale sont disponibles à l'adresse https://osf.io/ztcyb/. Public Significance Statement Contextual cues have been shown to automatically trigger adjustments in selective attention independent of awareness and intention. Here, we find that task relevance plays an important role in determining which context cues are used to direct attention. These findings contribute to a better understanding of how context-dependency might occur in more complex environments and more generally, how learning and memory processes enable flexible control over attention.