Abstract: The disclosure herein relates to methods and systems for generating an end-to-end de-smoking model for removing smoke present in a video. Conventional data-driven based de-smoking approaches are limited mainly due to lack of suitable training data. Further, the conventional data-driven based de-smoking approaches are not end-to-end for removing the smoke present in the video. The de-smoking model of the present disclosure is trained end-to-end with the use of synthesized smoky video frames that are obtained by source aware smoke synthesis approach. The end-to-end de-smoking model localize and remove the smoke present in the video, using dynamic properties of the smoke. Hence the end-to-end de-smoking model simultaneously identifies the regions affected with the smoke and performs the de-smoking with minimal artifacts. localized smoke removal and color restoration of a real-time video.

Abstract: The disclosure herein relates to methods and systems for localized smoke removal and color restoration of a real-time video. Conventional techniques apply the de-smoking process only on a single image, by finding the regions having the smoke, based on manual air-light estimation. In addition, regaining original colors of de-smoked image is quite challenging. The present disclosure herein solves the technical problems. In the first stage, video frames having the smoky and smoke-free video frames are identified, from the video received in the real-time. In the second stage, an air-light is estimated automatically using a combined feature map. An intermediate de-smoked video frame for each smoky video frame is generated based on the air-light using a de-smoking algorithm. In the third and the last stage, a smoke-free video reference frame is used to compensate for color distortions introduced by the de-smoking algorithm in the second stage.

Abstract: This disclosure relates to a system and method for attention-based surface crack segmentation. Existing methods do not efficiently handle the sub-problem of data imbalance and inaccurate predicted pixels are ignored. The present disclosure obtains a binary edge map by passing a m-channel image through an edge detection algorithm and concatenate the obtained binary edge map with a channel dimension to obtain a (m+1)-channel image. Feature maps are extracted from an encoder and a decoder by feeding the obtained (m+1)-channel image into a network, wherein the feature maps are convolved with an attention mask and merged in a fused network. The merged feature maps are up sampled and concatenated to obtain a final fused feature map. The final fused feature map is passed through a sigmoid activation function to obtain a probability map which is iteratively thresholded to obtain a binary predicted image. The binary image is indicative of crack pixels.

Abstract: This disclosure relates generally to method and system to calculate net carbon sequestration for agriculture using remote sensing data. Climate change is one of the factor in sustainable development on the earth and has sparked numerous initiatives to reduce earth's carbon footprint. The disclosed method processes remote sensing data comprising one or more input images indicating one or more characteristics of at least one agriculture crop of a geographical region. The method calculates a carbon footprint value of at least one agriculture crop by obtaining a plurality of carbon values associated with the geographical region. A net carbon flux of least one agriculture crop is calculated based on the carbon footprint value, a data maturity index, and a difficulty level. The method enables growers and carbon credit purchasers to correctly determine the carbon footprint of a geographical region to accurately predict the reduction of greenhouse gas emissions reducing environmental degradation.

Abstract: Technical challenge in unusual human activity detection task is to rightly identify only unexpected or unusual movements from constant regular movements present in a scene, with most techniques built on understanding that camera is static. However, ego view camera of mobile surveillance robot is in motion as robot navigates. Embodiments herein provide a method and system for anomalous activity detection for mobile surveillance robots by mimicking ‘Konio-Parvocellular-Magno’ cells of the human brain into a NN model, which are responsible for detecting slow, normal, and swift changes in perceived scenes. To detect anomalous activity, the static or normal movements of scene captured by ego view camera are identified as redundant information and only RoI is forwarded for further processing using the Optical flow and SSIM techniques. The NN model mimicking KPM is trained only on the RoI to detect normal or anomalous activity.

Abstract: State of art techniques performing image labeling of remotely sensed data are computation intensive, consume time and resources. A method and system for efficient retrieval of a target in an image in a collection of remotely sensed data is disclosed. Image scanning is performed efficiently, wherein only a small percentage of pixels from the entire image are scanned to identify the target. One or more samples are intelligently identified based on sample selection criteria and are scanned for detecting presence of the target based on cumulative evidence score Plurality of sampling approaches comprising active sampling, distributed sampling and hybrid sampling are disclosed that either detect and localize the target or perform image labeling indicating only presence of the target.

Abstract: Embodiments of the present disclosure provide a method and system for co-operative and cascaded inference on the edge device using an integrated Deep Learning (DL) model for object detection and localization, which comprises a strong classifier trained on largely available datasets and a weak localizer trained on scarcely available datasets, and work in coordination to first detect object (fire) in every input frame using the classifier, and then trigger a localizer only for the frames that are classified as fire frames. The classifier and the localizer of the integrated DL model are jointly trained using Multitask Learning approach. Works in literature hardly address the technical challenge of embedding such integrated DL model to be deployed on edge devices. The method provides an optimal hardware software partitioning approach for components or segments of the integrated DL model which achieves a tradeoff between latency and accuracy in object classification and localization.

Abstract: Embodiments herein provide a method and system for a hyperspectral artificial vision for machines. The system receives a hyperspectral signal of a target material as an input to a neural network model. The system initializes by selecting the number of primitive layers to be used. The system iteratively cycles through all training data (pixels) and updating weights for each unsuccessful material class prediction. Model with two primitives serves as baseline, after which the system adds another primitive layer and repeats the training procedure. The system keeps repeating these processes until obtains convergence. Where the system come to a halt, the system obtains the optimal number of primitives for the given materials. The generated new color pixel is used as a discriminator to aid in locating the target material. The new artificial color is a mixture of weighted chromatic primitives which are optimized for sensitivity/(Spectral Response Functions) SRFs.

Abstract: Embodiments herein provide a method and system for a vicarious calibration of optical data from satellite sensors for urban scene flat fields. Identifying test sites automatically in the urban scene helps in vicarious calibration or on-board calibration of the hyperspectral/multispectral image. An internal average relative reflectance is calculated to get a relative reflectance of the image. Band ratios for various pixels is determined to assess flatness of the spectrum. Flat field candidates are identified from the various pixels having average band ratio nearing zero and a morphological technique is applied to determine a flat field. Finally, the image is calibrated vicariously based on the determined flat field as a test site. The on-board calibration of the remote sensing image may lead to a faster way to get the reflectance image of the scene, with the help of the calibration constants.

Abstract: This disclosure relates generally to Synthetic Aperture Radar (SAR) reconstruction and finds wide application in remote sensing. Conventional approaches either involve huge computational requirement for processing or require specialized hardware along with many additional Radio Frequency (RF) components. The present disclosure provides two approaches for temporally sampling a received pulse compressed signal at two sub-sampling factors, wherein both methods involve frugal hardware implementation. Reconstruction approach of the art is based on the principle of difference ruler and is not suitable for SAR image reconstruction due to the large measurements and image dimensions. In accordance with the present disclosure, the reconstruction problem is framed as an inverse imaging problem by suitably using a forward model and employing an approach like Alternating Direction Method of Multipliers (ADMM) for solving this model which allows use of readily available Plug and Play (PnP) priors.

Abstract: The disclosure generally relates to scene graph generation. Scene graph captures rich semantic information of an image by representing objects and their relationships as nodes and edges of a graph and has several applications including image retrieval, action recognition, visual question answering, autonomous driving, robotics. However, to leverage scene graphs, computationally efficient scene graph generation methods are required, which is very challenging to generate due presence of a quadratic number of potential edges and computationally intensive/non-scalable techniques for detecting the relationship between each object pair using the traditional approach. The disclosure proposes a combination of edge proposal neural network and the Graph neural network with spatial message passing (GNN-SMP) along with several techniques including a feature extraction technique, object detection technique, un-labelled graph generation technique and a scene graph generation technique to generate scene graphs.

Abstract: This disclosure relates generally to identification and mitigation of bias while training deep learning models. Conventional methods do not provide effective methods for bias identification, and they require pre-defined concepts and rules for bias mitigation. The embodiments of the present disclosure train an auto-encoder to produce a generalized representation of an input image by decomposing into a set of latent embedding. The set of latent embedding are used to learn the shape and color concepts of the input image. The feature specialization is done by training an auto-encoder to reconstruct the input image using the shape embedding modulated by color embedding. To identify the bias, permutation invariant neural network is trained for classification task and attribution scores corresponding to each concept embedding are computed. The method also performs de-biasing the classifier by training it with a set of counterfactual images generated by modifying the latent embedding learned by the auto-encoder.

Abstract: Optical images in remote sensing are contaminated by cloud cover and bad weather conditions and are only available during the daytime. Whereas SAR images are completely cloud free, independent of weather conditions and can be acquired both during the day and at night. However, due to the speckle effect and side looking imaging mechanism of SAR images, they are not easily interpretable by untrained people. To address this issue, the present disclosure provides a method and system for LULC guided SAR visualization, wherein a GAN is trained to translate SAR images to optical images for visualization. A given SAR image is fed into a first generator of the GAN to obtain LULC map which is then concatenated with the SAR image and fed into a second generator of the GAN to generate an optical image. The LULC map provides semantic information required for generation of more realistic optical image.

Abstract: This disclosure relates generally to systems and methods for autonomous task composition of vision pipelines using an algorithm selection framework. The framework leverages transformer architecture along with deep reinforcement learning techniques to search an algorithmic space for unseen solution templates. In an embodiment, the present disclosure describes a two stage process of identifying the vision pipeline for a particular task. At first stage, a high-level sequence of the vision pipeline is provided by a symbolic planner to create the vision workflow. At second stage, suitable algorithms for each high-level task are selected. This is achieved by performing a graph search using a transformer architecture over an algorithmic space on each component of generated workflow. In order to make the system more robust, weights of embedding, key and query networks of a visual transformer are updated with a Deep Reinforcement Learning framework that uses Proximal Policy Optimization (PPO) as underlying algorithm.

Abstract: Conventional task planners assume that the task-plans provided are executable, hence these are not task-aware. Present disclosure alleviates the downward refinability assumption, that is, planning can be decomposed separate symbolic and continuous planning steps by introducing bi-level planning, a plan which is a series of actions that the robot needs to take to achieve the goal task is curated. Firstly, abstract symbolic actions are converted to continuous vectors and used therein to enable interaction with an environment. Images of objects placed in the environment are captured and concepts are learnt from the captured images and attributes of objects are detected. A hierarchical scene graph is generated from the concepts and attributes wherein the graph includes interpretable sub-symbolic representations and from these interpretable symbolic representations are obtained for identifying goal task.

Abstract: Assessing sustainability performance of an enterprise is a challenging task. Embodiments of present disclosure provide a method and system for SDG performance assessment of an enterprise with a conceptually simpler data model and processing pipeline. Enterprise data collected from hard and soft sensors is mapped to appropriate indicators of the SDGs. Further, a semantic network is constructed with nodes corresponding to each indicator and edges connecting nodes belonging to same SDG. Each node of the semantic network is further linked to a first layer of a neuro fuzzy network which calculates degree of impact of the indicator on Social, Economic and Environment values. Output of the first layer activates second layer of the neuro fuzzy network which determines BBV scores indicating whether the indicator is a burden, benefit, or vulnerability. The BBV scores are transformed to a colour space to generate a colour that indicates SDG performance of the enterprise.

Abstract: Current inspection processes employed for pipeline networks data acquisition aided with manually locating and recording defects/observations, thus leading labor intensive, prone to error and a time-consuming task thereby resulting in process inefficiencies. Embodiments of the present disclosure provide systems and methods for that leverage artificial intelligence/machine learning models and image processing techniques to automate log and data processing, reports and insights generation thereby reduce dependency on manual analysis, improve annual productivity of survey meterage and bring in process and cost efficiencies into overall asset health management for utilities, thereby enhancing accuracy in defect identification, analysis, classification thereof.

Abstract: This disclosure relates generally to method and system for live video streaming with integrated encoding and transmission semantics. The system receives a set of frames associated with a live video stream encoded to generate a set of data fragments using a reference encoder and a delta encoder. Transmitter unit of the live video streaming protocol transmits each packet of the set of full frames and the set of delta frames in sequence with a payload specific header based on a packet mode. Further, the receiver unit receives each packet of the full frames and each packet of the delta frames based on the packet mode to reconstruct an original sequence from the foreground pixels by estimating a total number of packets expected at each frame interval and loss incurred in each packet of the set of full frames and the set of delta frames.

Abstract: State of the art systems used for airport automation and data processing may be prone to data security related issues, as unauthorized personal may gain entry to sensitive data. The disclosure herein generally relates to airport management, and, more particularly, to a method and system for service authentication in an airport management network. The system uses a neural network to process a received service request and decides whether the service request is to be allowed or denied, based on a determined validity of the service request, role based access defined for a user requesting the service, a feature map data generated.

Abstract: This disclosure relates generally to method and system for live video streaming with integrated encoding and transmission semantics. The system receives a set of frames associated with a live video stream encoded to generate a set of data fragments using a reference encoder and a delta encoder. Transmitter unit of the live video streaming protocol transmits each packet of the set of full frames and the set of delta frames in sequence with a payload specific header based on a packet mode. Further, the receiver unit receives each packet of the full frames and each packet of the delta frames based on the packet mode to reconstruct an original sequence from the foreground pixels by estimating a total number of packets expected at each frame interval and loss incurred in each packet of the set of full frames and the set of delta frames.