Search

To travel securely on roads, traffic signs are very important. The main cause of road accidents is laxity in viewing the traffic signboards and analysing them wrongly. To avoid all these problems, the new system is implemented. This system helps in identifying the traffic sign and sends a voice command to the driver, so that the driver will take necessary actions to travel safely. The system is trained by using Convolutional Neural Network (CNN/ConvNet), helps in classification and recognisation of traffic sign images. To make the system more accurate, a particular dataset is taken and trained by defining a set of classes. The German Traffic Sign Recognition Benchmark dataset was used for classification of sign boards,it contains more than 40 classes and more than 50000 images of traffic signs. After detecting the sign, a voice command is sent to notify the driver and aware him/her to follow the necessary rules on the road. The main aim of this system is safe and secure travelling and the lives of the people. [ABSTRACT FROM AUTHOR]

India yields many varieties of crops. Among them tomato crop is the vital staple in Indian market. Many farmers produce tomato crop cross India. Many types of diseases will affect the crop due to many reasons. So, it is important to prevent the crop from being affected by those diseases and also curing the affected portion with suitable techniques. Our model is to predict the tomato leaf's health and also identify type of the disease it is affected with. The proposed model targets to reduce the amount of chemical fertilized used without having proper knowledge of the disease affected. Also help the cultivator to grow the tomato plants properly. This model is developed using Deep Convolutional Neural Network (CNN) model called TOMLeNet and Adam optimizer is used to predict the disease in the early stage and to avoid loss to the farmer. Neural network models make use of automatic feature extraction for the classification of the input image into their respective disease classes. The proposed model uses the tomato leaf disease data set publicly available in Kaggle and model generated an average accuracy of 97.42% using the features provide and even in unfavorable conditions. [ABSTRACT FROM AUTHOR]