Google parses everything, what it can parse
Since some years Google reads image metadata, so called Exif - that is describing image data, which is written "into" the image, namely into JPEG and TIFF (gif and png are not supported). This Exif metadata contains describing fields filled with info about several aspects of image charachter, nature, topic, author and quality. It is clear, that this bunch of data requires some volume. I tested it on this site's logo image: without any written Exif data the image weight is 4kb, with all filled fields (with one word "test") image weight grows on 6kb. Also befor you publish any image, firstly better get rid of any written Exif data to get the image size lower, specially if you don't plan to write any Exif into it. You can delete all Exif meta from a given image directly online, with this free online exif removing tool.
Step further into image SEOBut Exif format has some fields, which are extremely important for filling, if we are seriously on SEO. There are e.g.: artist, copyright, document_name, image_description, maker_note, subject_area, subject_location, user_comment, xp_author, xp_comment, xp_subject, xp_title. Each field might contain 2kb of data, also ca 1 text page;) Enough to feed Google, nope?
To write Exif into your images there is a real bunch of tools: free, commercial, online, offline. Here is the free online Exif editor of my choice. Add an image, fill in metadata into fields, upload image for processing, then download image with written Exif metadata, simple enough. I haven't use any offline Exif editor ever - you will search it alone. But i'm sure: you will manage it;) If you are interesting in more advices, i recommend you this article with 10 valueable advanced image optimization tips.