Tessaract gives poor results on Binary Image - c++

Question

I am trying to detect Alphabets in a picture using the tesseract library, which comes as an extra module for OpenCv3.
After a lot of preprocessing, I was able to get a very decent enough picture of the Alphabet.
However the tesserct library fails to recognise it at all.
Here is what I have written so far:
#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <opencv2/text/ocr.hpp>
#include <opencv2/imgproc/imgproc.hpp>
#include <iostream>
#include <stdlib.h>
using namespace cv;
using namespace std;
using namespace cv::text;
Mat img=imread("result.jpg",0);
if(img.empty()){
cout<<"Image not loaded!!\n";
return 1;
}
string output_text;
vector<Rect> detected;
vector<string> list;
vector<float> conf;
Ptr<OCRTesseract> ocr = OCRTesseract::create();
ocr->run(img,output_text,&detected,&list,&conf,0);
for(int i=0; i<detected.size(); i++){
Scalar color = Scalar(255,255,150);
rectangle( img, detected[i].tl(), detected[i].br(), color, 2, 8, 0 );
}
//cout<<"x="<<detected[0].x<<" y="<<detected[0].y
cout<<"Text found is:\n"<<output_text<<endl;
imshow("My",img);
waitKey(0);
return 0;
People have said that Tesseract is my best option. However I am not getting any result with the same.
Here is the link to Tesseract Library Documentation for opencv3.
http://docs.opencv.org/trunk/d7/ddc/classcv_1_1text_1_1OCRTesseract.html#gsc.tab=0
Also, is there any difference between using the native Tesseract Lib vs OpenCV version ?