To generate a pixel array from a PNG, JPG or BMP image, use the [Online image converter tool](https://lvgl.io/tools/imageconverter) and set the converted image with its pointer: `lv_img_set_src(img1, &converted_img_var);`
To use external files, you also need to convert the image files using the online converter tool but now you should select the binary output format.
You also need to use LVGL's file system module and register a driver with some functions for the basic file operation. Go to the [File system](/overview/file-system) to learn more.
You can set a symbol similarly to [Labels](/widgets/core/label). In this case, the image will be rendered as text according to the *font* specified in the style. It enables to use of light-weighted mono-color
Images and labels are sometimes used to convey the same thing. For example, to describe what a button does.
Therefore, images and labels are somewhat interchangeable, that is the images can display texts by using `LV_SYMBOL_DUMMY` as the prefix of the text. For example, `lv_img_set_src(img, LV_SYMBOL_DUMMY "Some text")`.
A color can be mixed to every pixel of an image with a given intensity.
It is very useful to show different states (checked, inactive, pressed, etc.) of an image without storing more versions of the same image.
This feature can be enabled in the style by setting `img_recolor_opa` between `LV_OPA_TRANSP` (no recolor, value: 0) and `LV_OPA_COVER` (full recolor, value: 255).
Is the width or height of the image object is set to `LV_SIZE_CONTENT` the obejct's size will be set according to the size of image source in the respective direction.
With `lv_img_set_offset_x(img, x_ofs)` and `lv_img_set_offset_y(img, y_ofs)`, you can add some offset to the displayed image.
It is useful if the object size is smaller than the image source size.
Using the offset parameter a [Texture atlas](https://en.wikipedia.org/wiki/Texture_atlas) or a "running image" effect can be created by [Animating](/overview/animation) the x or y offset.
## Transformations
Using the `lv_img_set_zoom(img, factor)` the images will be zoomed. Set `factor` to `256` or `LV_IMG_ZOOM_NONE` to disable zooming.
A larger value enlarges the images (e.g. `512` double size), a smaller value shrinks it (e.g. `128` half size).
Fractional scale works as well. E.g. `281` for 10% enlargement.
To rotate the image use `lv_img_set_angle(img, angle)`. Angle has 0.1 degree precision, so for 45.8° set 458.
By default, the pivot point of the rotation is the center of the image. It can be changed with `lv_img_set_pivot(img, pivot_x, pivot_y)`. `0;0` is the top left corner.
The quality of the transformation can be adjusted with `lv_img_set_antialias(img, true/false)`. With enabled anti-aliasing the transformations has a higher quality but they are slower.
The transformations require the whole image to be available. Therefore indexed images (`LV_IMG_CF_INDEXED_...`), alpha only images (`LV_IMG_CF_ALPHA_...`) or images from files can not be transformed.
In other words transformations work only on true color images stored as C array, or if a custom [Image decoder](/overview/images#image-edecoder) returns the whole image.
Note that, the real coordinates of image object won't change during transformation. That is `lv_obj_get_width/height/x/y()` will returned the original, non-zoomed coordinates.