1、Using AES-EME2 algorithm. It seemed to be more secure.
Don't know anything about this but using a "new" algorithm is risky. Not to be insulting but "seems more secure"? You would really want to implement a new algorithm on a hunch? Twofish and Serpent are already in there if you don't trust AES.
2、One can freely choose 'offset' value to build a real hidden volume or partition. Without the right offset value, one can't decrypt a disk even he has the right password.
Maybe but you would have to remember the offset or have it as part of a keyfile of sorts.
3、No hashing result of the password store in the header, just 'decrypt' the volume or partition even when input a bad password, and get a bad driver. Because there's no way to judge whether a password is right or wrong, this could stop brute force attacks.
There is no real benefit to this. An attacker can still look for known structures in the file system that are predictable. It would only slow attacks by a few percent. You would still have to verify this each time a volume is mounted, unless you are one of those people that has never mistyped a password!
4、Each sector on the disk is encrypted with a different master key. So far most encryption tools encrypt the whole volume or partition using a single master key or only a few keys.
Each data block is encrypted with a different key. The master key is combined with the block location to "tweak" the key. See XTS mode of operation. What you are proposing is a One Time Pad where the key is as large as the data it protects. OTP is proven unbreakable but impractical for real use.