Aditi Raut - Academia.edu (original) (raw)

Papers by Aditi Raut

Research paper thumbnail of Suspicious Activity Detection Using Machine Learning

International Journal for Research in Applied Science and Engineering Technology, May 31, 2023

Various technologies have been utilized to implement the safety of life and property by installin... more Various technologies have been utilized to implement the safety of life and property by installing high quality CCTV cameras. It is not possible to manually monitor each and every moment activity. Furthermore, in practical scenario the most unpredictable one is human behaviour and it is very difficult to find whether it is suspicious or normal. In this work the notion of CNN is used to detect suspicious or normal activity in an environment, and a system is proposed that sends an alert message to the similarity authority, in case of predicting a suspicious activity. It's worth noting that the effectiveness of a suspicious activity detection system relies on the quality of the training data, the architecture of the Machine Learning model, and the deployment environment. Ongoing monitoring, regular updates, and continuous improvement are important for maintaining the system's accuracy and adapting it to new and emerging types of suspicious activities.

Research paper thumbnail of Automatic Image and Video Colourisation using Deep Learning

2018 International Conference on Smart City and Emerging Technology (ICSCET), 2018

A pivotal area of research among the machine learning and computer vision communities is the Colo... more A pivotal area of research among the machine learning and computer vision communities is the Colourisation of monochrome/black and white images. Colourisation is the computer-assisted process of adding colour to a greyscale image/movie. Traditionally, this process required significant user interaction, in the form of placing numerous colour scribbles, looking at related images, and performing segmentation. However, with advancements in technology, automated Colourisation systems have been created. Apart from the aesthetic appeal, such capability has broad practical applications ranging from video restoration to image enhancement for improved comprehensibility. However, these current systems face some major challenges, such as - Colour inconsistency within individual objects, under/over-saturation, and green tones in bright environments. The main purpose here is to eliminate the issues faced by current systems, and at the same time, efficiently colourise Black & White images and videos.

Research paper thumbnail of Automatic Image and Video Colourisation using Deep Learning

2018 International Conference on Smart City and Emerging Technology (ICSCET)

A pivotal area of research among the machine learning and computer vision communities is the Colo... more A pivotal area of research among the machine learning and computer vision communities is the Colourisation of monochrome/black and white images. Colourisation is the computer-assisted process of adding colour to a greyscale image/movie. Traditionally, this process required significant user interaction, in the form of placing numerous colour scribbles, looking at related images, and performing segmentation. However, with advancements in technology, automated Colourisation systems have been created. Apart from the aesthetic appeal, such capability has broad practical applications ranging from video restoration to image enhancement for improved comprehensibility. However, these current systems face some major challenges, such as - Colour inconsistency within individual objects, under/over-saturation, and green tones in bright environments. The main purpose here is to eliminate the issues faced by current systems, and at the same time, efficiently colourise Black & White images and videos.

Research paper thumbnail of Handwritten Text Image Recognition Using Feature Extraction

In this paper,a new method is proposed which recognizes English handwritten text based on its fea... more In this paper,a new method is proposed which recognizes English handwritten text based on its features. This framework consists of a formal model definition and the algorithm for recognition. In pre-processing stage, determinant value makes recognition process feasible for recognizing given text from the dataset. The determinant value produces the feature, which is obtained by the division of the image into blocks. Later with the help of chain code further recognition is done. The output text file is matched with the one in the database to check the similarity.

Research paper thumbnail of Web Logs Analysis for Finding Brand Status

IOSR Journal of Computer Engineering, 2014

Due to rapid development of the web there has been vast increase in the user generated contents a... more Due to rapid development of the web there has been vast increase in the user generated contents available in the form of blogs ,product reviews sites, web-forums and online social networks etc. Such reviews are very useful to the companies, as they contain valuable information about what aspects of the product, are driving the sales up or down.But most of the available content is in the form of unstructured data from where extraction of information is a challenging task. In order to find potential risk, it is essential for companies to gather and analyze information about competitor products and strategies. Companies use opinion mining techniques to track customer's response, in order to efficiently market their products, identify new opportunities, weaknesses and manage their reputations. In this paper, brand status model is proposed based on the various parameters which affect brands reputation. Also, the products belonging to different brands will be compared on the basis of predefined aspects. This model uses aspect based sentiment analysis technique to find sentiments about product's features and brand.

Research paper thumbnail of Suspicious Activity Detection Using Machine Learning

International Journal for Research in Applied Science and Engineering Technology, May 31, 2023

Various technologies have been utilized to implement the safety of life and property by installin... more Various technologies have been utilized to implement the safety of life and property by installing high quality CCTV cameras. It is not possible to manually monitor each and every moment activity. Furthermore, in practical scenario the most unpredictable one is human behaviour and it is very difficult to find whether it is suspicious or normal. In this work the notion of CNN is used to detect suspicious or normal activity in an environment, and a system is proposed that sends an alert message to the similarity authority, in case of predicting a suspicious activity. It's worth noting that the effectiveness of a suspicious activity detection system relies on the quality of the training data, the architecture of the Machine Learning model, and the deployment environment. Ongoing monitoring, regular updates, and continuous improvement are important for maintaining the system's accuracy and adapting it to new and emerging types of suspicious activities.

Research paper thumbnail of Automatic Image and Video Colourisation using Deep Learning

2018 International Conference on Smart City and Emerging Technology (ICSCET), 2018

A pivotal area of research among the machine learning and computer vision communities is the Colo... more A pivotal area of research among the machine learning and computer vision communities is the Colourisation of monochrome/black and white images. Colourisation is the computer-assisted process of adding colour to a greyscale image/movie. Traditionally, this process required significant user interaction, in the form of placing numerous colour scribbles, looking at related images, and performing segmentation. However, with advancements in technology, automated Colourisation systems have been created. Apart from the aesthetic appeal, such capability has broad practical applications ranging from video restoration to image enhancement for improved comprehensibility. However, these current systems face some major challenges, such as - Colour inconsistency within individual objects, under/over-saturation, and green tones in bright environments. The main purpose here is to eliminate the issues faced by current systems, and at the same time, efficiently colourise Black & White images and videos.

Research paper thumbnail of Automatic Image and Video Colourisation using Deep Learning

2018 International Conference on Smart City and Emerging Technology (ICSCET)

A pivotal area of research among the machine learning and computer vision communities is the Colo... more A pivotal area of research among the machine learning and computer vision communities is the Colourisation of monochrome/black and white images. Colourisation is the computer-assisted process of adding colour to a greyscale image/movie. Traditionally, this process required significant user interaction, in the form of placing numerous colour scribbles, looking at related images, and performing segmentation. However, with advancements in technology, automated Colourisation systems have been created. Apart from the aesthetic appeal, such capability has broad practical applications ranging from video restoration to image enhancement for improved comprehensibility. However, these current systems face some major challenges, such as - Colour inconsistency within individual objects, under/over-saturation, and green tones in bright environments. The main purpose here is to eliminate the issues faced by current systems, and at the same time, efficiently colourise Black & White images and videos.

Research paper thumbnail of Handwritten Text Image Recognition Using Feature Extraction

In this paper,a new method is proposed which recognizes English handwritten text based on its fea... more In this paper,a new method is proposed which recognizes English handwritten text based on its features. This framework consists of a formal model definition and the algorithm for recognition. In pre-processing stage, determinant value makes recognition process feasible for recognizing given text from the dataset. The determinant value produces the feature, which is obtained by the division of the image into blocks. Later with the help of chain code further recognition is done. The output text file is matched with the one in the database to check the similarity.

Research paper thumbnail of Web Logs Analysis for Finding Brand Status

IOSR Journal of Computer Engineering, 2014

Due to rapid development of the web there has been vast increase in the user generated contents a... more Due to rapid development of the web there has been vast increase in the user generated contents available in the form of blogs ,product reviews sites, web-forums and online social networks etc. Such reviews are very useful to the companies, as they contain valuable information about what aspects of the product, are driving the sales up or down.But most of the available content is in the form of unstructured data from where extraction of information is a challenging task. In order to find potential risk, it is essential for companies to gather and analyze information about competitor products and strategies. Companies use opinion mining techniques to track customer's response, in order to efficiently market their products, identify new opportunities, weaknesses and manage their reputations. In this paper, brand status model is proposed based on the various parameters which affect brands reputation. Also, the products belonging to different brands will be compared on the basis of predefined aspects. This model uses aspect based sentiment analysis technique to find sentiments about product's features and brand.