A
A
altair862018-08-08 11:27:37
.NET
altair86, 2018-08-08 11:27:37

How to convert a jpg/png image to the smallest size?

There is a photo of the interface of the web page - forms in fact - where there is a set of the most primitive elements: segments, text, rectangles. However, the photo of this page is considerable: 30-40 kB with significant compression. If the compression is increased, then it ceases to be clear what is shown, it is impossible to read the text.
At the same time, many gigabytes of data files are typed.
Is it possible to convert a jpg/png image to the smallest size? If so, how? Is it possible to convert a raster image to a vector to get better results? If so, how?

Answer the question

In order to leave comments, you need to log in

2 answer(s)
R
Roman Mirilaczvili, 2018-08-08
@altair86

You can try formats:
No web browser support!
man cpaldjvu

NAME
cpaldjvu - DjVuDocument encoder for low-color images.
SYNOPSIS
cpaldjvu [options] inputppmfile outputdjvufile
DESCRIPTION
Program cpaldjvu is a DjVuDocument encoder for images containing few colors. It performs best on images containing large solid color areas such as screen dumps. Compression ratios on such
images can be much higher than those achieved by GIF or PNG compression.
This program works by first reducing the number of distinct colors to a small specified value using a simple color quantization algorithm. The dominant color is encoded into the background
layer. The other colors are encoded into the foreground layer.
155487 screenshot-test1.png (original image of Toaster)
49461 screenshot-test16.djvu (limited color palette to 16)
61675 screenshot-test32.djvu (... 32) - good quality/size balance!
94641 screenshot-test256.djvu (... 256)
Dimensions in bytes.
  • Convert using OCR to SVG - only theoretically. But in practice it gives poor results even on ABBYY OCR at low DPI (regular screenshot).
    It is also possible to store a picture in poor resolution and store OCR recognized text
  • G
    Griboks, 2018-08-08
    @Griboks

    web page interface

    Great, so any element can be encoded with a code. for example
    name=page1
    title=Привет, мир!
    text=Моя первая страница
    next=page2,page8

    When viewing the desired "image" (state), we generate a static page with the properties described by the code.

    Didn't find what you were looking for?

    Ask your question

    Ask a Question

    731 491 924 answers to any question