Description
Millimeter Wave Doughnut Slot MIMO Antenna
Description
Does the project present logo characteristics analysis using image processing ? ? or techniques for automated vision system used in the agricultural fields.? In agriculture research, automatic logo characteristics detection is essential in monitoring large fields of crops, and thus automatically detect symptoms of logo characteristics as soon as they appear on plant leaves. The proposed decision-making system utilizes image content characterization and a supervised classifier type of neural network. Image processing techniques for this kind of decision analysis involve preprocessing feature extraction and a classification stage. At Processing, an input image will be resized and region of interest selection performed if needed. Here, color and texture features are extracted from the input for network training and classification. Color features like mean, the standard deviation of HSV color space, and texture features like energy, contrast, homogeneity, and correlation. The system will be used to classify the test images automatically to decide logo characteristics. For this approach, the automatic classifier NN is used for classification based on learning with some training samples of that same category. This network uses the tangent sigmoid function as the kernel function. Finally, the simulated result shows that the used network classifier provides minimum error during training and better accuracy in classification. Millimeter Wave Doughnut Slot MIMO Antenna
Existing method
Most of the existing methods are based on connected component analysis (CCA) and don’t work in real-time due to the many customized processing steps involved. Hence, considering Deep CNNs is quite intuitive, given the recent success in basic computer vision problems.
PROPOSED method
This paper investigates the possibility of modeling the signature and logo detection task as an end-to-end object detection problem
In this paper, we proposed to use the state-of-the-art Deep Convolutional Neural Networks (CNN) for detecting signatures and logos from scanned documents. Specifically, we analyze the potential of Faster-RCNN [17] and YOLOv2 [24] for the detection of the areas of interest and adapt them to the document retrieval problem. Four different network architectures namely ZF [18], VGG16 [19], VGG16-M-1024 [19], and Yolov2 [24] are used in this study. The primary intention is to explore and model the signature and logo detection task into a standard object detection problem. Additionally, real-time detection of signatures and logos in a single pipeline, makes it more applicable to document retrieval.
Methodologies
- Preprocessing
- Color Space Conversion
- Color and Texture Features Extraction
- NN classifier
Drawback
- A major drawback of this technique is that a priori information about the location of the signature is assumed. Ahmed et al. [1] proposed a Speeded Up Robust Features (SURF) based approach for signature segmentation from document images
Application
- Real-time detection of signatures and logos in a single pipeline makes it more applicable to document retrieval.
SOFTWARE REQUIREMENT
PYTHON ABOVE 3.0
Customer Reviews
There are no reviews yet.