Minor Research Project


Mrinmoyee Bhattacharya
Department of Computer Science


Determining a best route in highly developed complex transportation networks is not a trivial task, especially for those who are unfamiliar with the local transportation system. To assist the mobility of people by taking advantage of the multimodal transportation infrastructure is the main goal of intelligent multimodal navigation services. Multimodal route planning that aims to find an optimal route between the source and the target of a trip while utilizing several transportation modes including car driving, public transportation, cycling, walking, etc. is essential to intelligent multimodal navigation services. Although the task originates from the field of transportation, it can be abstracted as a general form independent of the domain-specific details on the underlying data model and algorithms. This research work is therefore dedicated to a general approach of modelling the multimodal network data and performing optimal path queries on it. The approach is approved in the application field of urban transportation. The bottleneck in the development of a multimodal route planning service is reflected in two aspects: one is the lack of a high-quality dataset; the other is the lack of effective modelling and path-finding approach. With an integrated navigation dataset produced from an automated data-matching process as the desirable test bed, this research is focused on the second aspect.

The weighted digraph structure can well represent the fundamental static networks. For each mode, there is one corresponding mode graph. These graphs constitute the Multimodal Graph Set as a key component of the overall multimodal network data model. In comparison with the traditional mono-modal problem, another key component necessary in the modelling of multimodal route-planning problem is mode-switching actions. In this work, such actions are described by Switch Points which are somewhat analog to plugs and sockets between different mode graphs. Consequently, it is possible to plug-and-play a Multimodal Graph Set by means of Switch Points.

On the basis of the multimodal network data model, the multimodal route-planning problem is categorized into two types and formalized as the multimodal shortest path problem on the Plug-and-Play Multimodal Graph Set. The first type where the mode sequence is given in the input is described as to find a shortest path from a given source to a destination across the modes in the sequence one after another. This type of problem can be solved within a general algorithmic framework. For the second type where the mode sequence cannot be determined beforehand, the multimodal path-finding algorithm can make good use of the traditional mono-modal shortest path algorithms together with the SCM-P LUG operation. It turns out that the solutions for these two types of problem are equivalent if the input mode list for the first type is transformed into its matrix expression. When applying the general multimodal route-planning approach to a specific application domain, a rule-based inferring process is necessary to determine whether a mode sequence is reasonable or not. Performance evaluations on the integrated navigation dataset have verified the efficiency I of the proposed approach. A web-based prototype system demonstrates the whole workflow of the multimodal route-planning function which is missing in any other existing systems. Case studies based on the prototype system show that all feasible routing plans and the corresponding optimal paths can be automaticallycreated for users who just need to tell the system about their preferences on the usage of transportation modes.


The Purpose of this multimodal route planning was to provide the traveller with an optimal, feasible and personalized routes between the source and destination, which may include public and private transportation modes. This project proposed an improved algorithm for route planning in multimodal, multicriteria environment. The various operations like crossover, mutation etc. have been redefined. This approach implements a free combination of travel modes according to their various individual needs, which has little manipulation and more of intelligence. This current multiobjective optimization is useful and it consumes less running time.


Principal Investigator:
Dr. M.N.Nachappa
Department Of Computer Science

I. Introduction to multimodal biometrics

Biometrics refers to the physiological or behavioral characteristics of a person to authenticate his/her identity. The increasing demand of enhanced security systems has led to an unprecedented interest in biometric based person authentication system. Biometric systems based on single source of information are called unimodal systems. Although some unimodal systems have got considerable improvement in reliability and accuracy, they often suffer from enrolment problems due to non-universal biometrics traits, susceptibility to biometric spoofing or insufficient accuracy caused by noisy data.

Hence, single biometric may not be able to achieve the desired performance requirement in real world applications. One of the methods to overcome these problems is to make use of multimodal biometric authentication systems, which combine information from multiple modalities to arrive at a decision. Studies have demonstrated that multimodal biometric systems can achieve better performance compared with unimodal systems.

This research proposes the study of multimodal biometrics for plant species identification. The different fusion techniques of multimodal biometrics will be studied.

II. Multimodal biometrics

The term “multimodal” is used to combine two or more different biometric sources of a plant sensed by different sensors. Two different properties of the same biometric can also be combined. In orthogonal multimodal biometrics, different biometrics are involved with little or no interaction between the individual biometric whereas independent multimodal biometrics processes individual biometric independently. Orthogonal biometrics are processed independently by necessity but when the biometric source is the same and different properties are sensed, then the processing may be independent, but there is at least the potential for gains in performance through collaborative processing. In collaborative multimodal biometrics the processing of one biometric is influenced by the result of another biometric.

A generic biometric system has sensor module to capture the trait, feature extraction module to process the data to extract a feature set that yields compact representation of the trait, classifier module to compare the extracted feature set with reference database to generate matching scores and decision module to determine an identity or validate a claimed identity.

In multimodal biometric system information reconciliation can occur at the data or feature level, at the match score level generated by multiple classifiers pertaining to different modalities and at the decision level.

III. Objectives of the study:

As mentioned above, the goal of this research is to develop a Multi Algorithm biometric system to identify plants using leaf patterns beyond overall leaf shape alone, which prior research has shown to be insufficient for classification.
There are three main objectives.
Analysis of existing pattern recognition techniques to identify the techniques and algorithms that can be used to detect leaf contours accurately, to represent the leaf margins preserving both global and local features and to match the leaf shapes efficiently.
Propose and develop a method to separate and to quantify the singularities on the leaf contour. Singularities of a leaf contour are the shape of the entire leaf, shape of leaf apex, shape of leaf margin and shape of leaf base.
Identify and extract the other parameters from the leaf image that can be used to differentiate leaves. These parameters will be based on leaf size, texture and color.
Apply the obtained results to different algorithms and sample the results to create an accurate method of identification.

IV. Conclusions

Although biometrics is becoming an integral part of the identity management systems, current biometric systems do not have 100% accuracy. Some of the factors that impact the accuracy of biometric systems include noisy input, non-universality, lack of invariant representation and non-distinctiveness. Further, biometric systems are also vulnerable to security attacks. A biometric system that integrates multiple cues can overcome some of these limitations and achieve better performance. Extensive research work has been done to identify better methods to combine the information obtained from multiple sources. It is difficult to perform information fusion at the early stages of processing (sensor and feature levels). In some cases, fusion at the sensor and feature levels may not even be possible. Fusion at the decision level is too simplistic due to the limited information content available at this level. Therefore, researchers have generally preferred integration at the matching score level which offers the best compromise between information content and ease in fusion.

We have carried out a detailed evaluation of the various techniques that have been proposed in literature in terms of their efficiency and robustness. First, we studied the impact of the different schemes on the performance of the multimodal biometric system. Our analysis shows that certain techniques are efficient and provide a good recognition performance. However, we also observed that they are not robust if the training data used for estimating the normalization parameters contains outliers.

We are proposing two different algorithms both neural based which will help in the identification process in a more efficient manner.

In the first algorithm, we proposed and implemented of leaf recognition system based on the leaf vein and shape for plant classification. We extract main vein from the input image, and leaf direction is determined using projection histograms of extracted main vein image. We were extracted twenty-one leaf features as distance, FFT, and convex hull for the leaf recognition.

In the experimental results, the proposed leaf recognition system showed a performance of 97.19%. From the experimental results, we can confirm that the recognition rate of the proposed advanced leaf recognition system was better than that of the existing leaf recognition system.

In the second algorithm we introduces a neural network approach for plant leaf recognition. The computer can automatically classify 32 kinds of plants via the leaf images loaded from digital cameras or scanners. PNN is adopted for it has fast speed on training and simple structure. 12 features are extracted and processed by PCA to form the input vector of PNN. Experimental result indicates that our algorithm is workable with an accuracy greater than 90% on 32 kinds of plants. Compared with other methods, this algorithm is fast in execution, efficient in recognition and easy in implementation.


» Founder
» College Anthem
» College Emblem and Motto
» Patron of the College
» Administration
» Disclosures


» Sciences
» Social Sciences
» Languages
» Physical Education


» Event Details
» SJC Handbook 2014 - 15
» Prospectus 2014- 15


» Research Overview
» Research in Chemistry
» Research in Botany
» Research in English

PB 27094, 36 Lalbagh Road
Karnataka, INDIA

Tel: +91 - 080 2227 4079
080 2221 1429
Email: desk@sjc.ac.in