Maybe these sources could help

http://www.ncbi.nlm.nih.gov/pubmed/17614915

http://www.cabdirect.org/abstracts/19831620409.html

I'm using one of those provided methods as well

We hybridizing crops like cassava you may have pigmented and nonpigmented strains.

Generally the nonpigmented plants are due to recessive genes. The highly colored condition such as purple plants are due to a dominant gene for anthocyanin production.

Use you nonpigmented strain as females and the colored strains as males in your crosses. The idea is the any seed from noncolored female having a seedling which is colored is a successful cross. The non pigmented female strain is demasculated before stigmas are receptive. Pollen ready purple plants are used as an anther pollen source and the tweezer used to collect pollen which is brushed on the stigma of the female when it is a receptive maturity. 

Breeders use a pod to row technique to evaluate and follow in a progeny breeding scheme with evaluation and culling of unwanted reactions and types.