Multi-Channel CNN-based Object Detection for Enhanced Situation Awareness

Multi-Channel CNN-based Object Detection for Enhanced Situation Awareness

Shuo Liu

Shuo Liu

Kelowna, British Columbia

We proposed a novel object detection framework using image fusion and convolutional neural networks (CNNs) for the military scenario.

Artificial Intelligence, Robotics

  • 0 Collaborators

  • 0 Followers

    Follow

Description

Object Detection is critical for automatic military operations. However, the performance of current object detection algorithms is deficient in terms of the requirements in military scenarios. This is mainly because the object presence is hard to detect due to the indistinguishable appearance and dramatic changes of object's size which is determined by the distance to the detection sensors. Recent advances in deep learning have achieved promising results in many challenging tasks. The state-of-the-art in object detection is represented by convolutional neural networks (CNNs), such as the fast R-CNN algorithm. These CNN-based methods improve the detection performance significantly on several public generic object detection datasets. However, their performance on detecting small objects or undistinguishable objects in visible spectrum images is still insufficient. In this study, we propose a novel detection algorithm for military objects by fusing multiple images using CNNs. We combine spatial, temporal and thermal information by generating a three-channel image, and they will be fused as CNN feature maps in an unsupervised manner. The backbone of our object detection framework is from the fast R-CNN algorithm, and we utilize cross-domain transfer learning technique to fine-tune the CNN model on generated multi-channel images. In the experiments, we validated the proposed method with the images from SENSIAC (Military Sensing Information Analysis Centre) database and compared it with the state-of-the-art. The experimental results demonstrated the effectiveness of the proposed method on both accuracy and computational efficiency.

Gallery

Default user avatar 57012e2942

SAI SRI RAM N. updated status

Default user avatar 57012e2942

SAI SRI RAM NIDAMANURI

Hi guys,this is sriram pursuing B.tech 3rd year 2nd semester in vijayawada.I am presently working on a project which requires machine learning,convolutional neural networks,image processing,python,php and other languages (used for building an website).The main motto of my project includes detecting the objects from a given image.This is pretty hard project for me but more interesting one because this project seperates me from others.I have a team with one member.I asked my teammate to learn numpy,scipy and pandas libraries .I have already installed numpy,scipy , pandasand other libraries without installing anaconda.I am getting many confusions in this project.But finally i have got a better overview about it and my team is currently working on it.I surely believe that my team can complete the project.Finally i have completed some courses like machine learning and presently preparing numpy,scipy,pandas,convolutional neural networks.

Medium profile pic

Bishaw S. created project Visual data parsing for automated reasoning expert system

Medium c095353a d87e 4ba5 bc2a cba7ef363fb3

Visual data parsing for automated reasoning expert system

Extract the data available in the input image/visual and process it to find the close search result available for the related terms. when a input visual containing the sentence "where is Mt. Everest ?" is given the output will be "Nepal", when a input visual containing the sentence "how to get to pokhara from kathmandu" the output will be google maps result with route, when the input visual containing "Gautam Buddha" is given the output will be wikipedia result about Buddha. the system will be able to

Default user avatar 57012e2942

muthu s. updated status

Default user avatar 57012e2942

muthu samy

RFP, Proactive pitch in , blockchain,IOT and Big Data sales pitch and strategies, engage with customers and provide solution design, consulting and recommendations. Provide consulting to customers in identifying Blockchain , IOT and Big Data use cases and then guiding them towards implementation of use cases. Big Data strategic planning, technology roadmap, talent acquisitions and mentor team for cutting edge technology competitiveness such as Hadoop, HIVE, HBase, Cassandra, AWS, Spark, R, big data analytics, data visualizations. Establish and manage governance for the implementation process , related to design of the target organization, production performance and managing change,

Default user avatar 57012e2942

muthu s. updated status

Default user avatar 57012e2942

muthu samy

Working on Trident Techno Minds Business & Emerging Markets Pre-Planning and delivery team for block chain , Internet of things ,Big Data services, strategy engagements with our existing and new customers in Big Data space.

Medium deepme

carmine s. created project EdibleMush

Medium 2812c143 73fb 46ae 8115 2715fe6b7e26

EdibleMush

Currently Ai-based applications are available, but none of these offers a guarantee of truthfulness. Since in nature there are several chameleon mushrooms, the application may not be able to recognize it.

The idea of this project offers a more accurate solution for the mushroom recognition based on 3 different images to uniquely validate the mushroom: Cap, Stem and Gills of Mushroom.

The project involves the application of image recognition techniques based on deepLearning algorithms, for now trained with caffe framework.

Once the prototype environment for the workflow has been developed, specific test and validation datasets will be made for the most common mushrooms.

Furthermore, the use of Movidius in the prototyping phase is envisaged, which uses the training models realized with the help of the colfax cluster to do inference from the edges without the cloud.

See More

No users to show at the moment.

No users to show at the moment.