### Evaluation 
The evaluation of Task 1 (hard classification) is based on average accuracy. [Accuracy](https://en.wikipedia.org/wiki/Accuracy_and_precision) is  the fraction of correct classifications:

$$  Accuracy = \frac{correct\ classifications}{all\ classifications}$$

In addition, each participant is required to provide a short description of the method.

To check the applicability of the proposed systems for real-time application, applicants are also requested to report the processing time per document page. 

### Input and output format specifications

The input to the classifier is a document page image. The output is a single value (class) for TASK 1.

### Submission Format
**For every image in the test set**, the submission file should contain two columns:  `ImageID` and `SCRIPT TYPE,` as illustrated in the  "sample_submission.csv" file.  The file should contain a header and have the following format:
![](https://imgur.com/fcjSiZw.png)

**Figure 1: Submission format.**